WEBVTT

00:00:01.080 --> 00:00:05.080
Good afternoon everyone, it's a pleasure
to see you in

00:00:05.080 --> 00:00:09.080
person and to put faces to names and see
you at different

00:00:09.080 --> 00:00:13.080
heights, which we usually have more of
than we do, because

00:00:13.080 --> 00:00:17.080
most of you don't share, you don't open
the video and in

00:00:17.080 --> 00:00:23.000
that there's no difference, it's probably
your students and our students.

00:00:23.000 --> 00:00:27.000
But you know us and you've seen us,

00:00:27.000 --> 00:00:31.160
because we've already been in sessions.

00:00:31.400 --> 00:00:35.400
So today we are fortunate to have here in
person an event

00:00:35.400 --> 00:00:39.400
where we will give the final touch,
hopefully a golden

00:00:39.400 --> 00:00:43.040
one, to the Artificial Intelligence
course, with

00:00:43.040 --> 00:00:46.040
activities for all ages and subjects.

00:00:46.040 --> 00:00:47.800
Well, I won't introduce myself because
you already know me.

00:00:47.800 --> 00:00:51.720
I am Gregorio Roble, I am a professor
here at the Juan Carlos University.

00:00:52.000 --> 00:00:56.000
I welcome you and thank you for coming
here, I know that

00:00:56.000 --> 00:00:59.360
many of you have come from very, very far
away.

00:01:00.160 --> 00:01:04.160
We have teachers from every remote corner
of the

00:01:04.160 --> 00:01:08.160
Community of Madrid, and that makes this
4th edition

00:01:08.160 --> 00:01:11.880
of the course a source of satisfaction
for us.

00:01:12.120 --> 00:01:16.120
Well, and without further delay, I'm
going to give the

00:01:16.120 --> 00:01:20.120
floor to Beatriz Rodríguez Merchán, who
is the Director

00:01:20.120 --> 00:01:24.120
General of Bilingualism and Quality of
Education of

00:01:24.120 --> 00:01:28.120
the Community of Madrid, who has come
here and who is

00:01:28.120 --> 00:01:31.240
going to give us the official welcome.

00:01:32.080 --> 00:01:33.080
Thank you.

00:01:33.080 --> 00:01:35.280
Okay, thanks.

00:01:35.280 --> 00:01:36.120
Okay, that's enough.

00:01:36.120 --> 00:01:38.560
Thank you Bego, thank you very much.

00:01:38.560 --> 00:01:40.240
Well, good afternoon everyone.

00:01:40.240 --> 00:01:44.240
First of all, I want to thank all the
teachers,

00:01:44.240 --> 00:01:48.240
speakers, organizers, and all the
collaborators

00:01:48.240 --> 00:01:51.960
who make this 4th edition of this course
possible,

00:01:51.960 --> 00:01:54.960
which I believe is so important.

00:01:54.960 --> 00:01:58.960
An initiative promoted by the Community
of Madrid in collaboration

00:01:58.960 --> 00:02:02.960
with this wonderful university that
always supports us in many

00:02:02.960 --> 00:02:06.960
training courses and as I mentioned
Gregorio, I think it is important

00:02:06.960 --> 00:02:10.360
that behind the training of teachers
there are universities

00:02:10.360 --> 00:02:13.360
that endorse the content and the courses
that we do.

00:02:13.400 --> 00:02:17.400
This edition reflects something very
important: the enormous interest

00:02:17.400 --> 00:02:21.400
that currently exists in understanding
and applying artificial

00:02:21.400 --> 00:02:25.400
intelligence in the educational field.
The fact that you are here today,

00:02:25.400 --> 00:02:29.200
and all those who are not here but have
been in the other sessions, says a lot

00:02:29.200 --> 00:02:31.120
about your teaching work.

00:02:31.280 --> 00:02:35.280
As we were discussing earlier with the
organizers,

00:02:35.280 --> 00:02:39.280
we had 150 places available, but we have
received

00:02:39.280 --> 00:02:43.960
more than 530 applications, which is
always a sign of success.

00:02:43.960 --> 00:02:47.400
We would have loved to be able to attend
to those more than 500 requests for

00:02:47.400 --> 00:02:49.160
organizational reasons.

00:02:49.160 --> 00:02:52.960
It's not always possible, but it says a
lot about the interest this course

00:02:52.960 --> 00:02:55.960
generates among all teachers, so I thank
you for it.

00:02:56.160 --> 00:03:00.160
I believe that artificial intelligence
should not be understood

00:03:00.160 --> 00:03:04.160
solely as a technological tool, although
it is one, because it

00:03:04.160 --> 00:03:08.160
is also an opportunity to develop new
methodologies, to foster

00:03:08.160 --> 00:03:12.160
critical thinking above all, which I
believe is one of the main

00:03:12.160 --> 00:03:16.160
tasks we have as teachers with such a
powerful tool as this, and

00:03:16.160 --> 00:03:20.160
to personalize it so that it helps us
personalize learning and

00:03:20.160 --> 00:03:23.960
prepare our students for a world that
obviously changes much

00:03:23.960 --> 00:03:26.960
faster than we can sometimes adapt in
schools.

00:03:26.960 --> 00:03:31.440
But along with the opportunities, we must
also address challenges.

00:03:31.480 --> 00:03:35.480
That is why it is essential that both
teachers and students

00:03:35.480 --> 00:03:39.480
understand what artificial intelligence
really is, how it

00:03:39.480 --> 00:03:43.280
evolves, what its benefits are, but also,
and above all, what its risks are and

00:03:43.280 --> 00:03:45.200
where its limits may lie.

00:03:45.840 --> 00:03:49.840
Talking about artificial intelligence in
education, I believe,

00:03:49.840 --> 00:03:53.840
necessarily involves talking about ethics
in that artificial

00:03:53.840 --> 00:03:57.840
intelligence, privacy, critical thinking,
responsible use,

00:03:57.840 --> 00:04:01.960
and above all, teacher training, which is
why we are here today.

00:04:02.280 --> 00:04:06.280
We need teachers like you who are
prepared not only

00:04:06.280 --> 00:04:10.280
to use these artificial intelligence
tools, but

00:04:10.280 --> 00:04:14.280
to teach your students to live with them
in a natural,

00:04:14.280 --> 00:04:18.040
but also safe, critical and conscious
way.

00:04:18.600 --> 00:04:22.600
Because educating today is not just about
knowledge, which we all

00:04:22.600 --> 00:04:26.680
know how to do, but also about helping to
interpret the world around us.

00:04:26.720 --> 00:04:30.640
So, well, I don't want to take up any
more of your time from these

00:04:30.640 --> 00:04:33.640
wonderful presentations that we're going
to have next.

00:04:33.640 --> 00:04:38.000
I just want to wish you a good session.

00:04:38.040 --> 00:04:41.680
Thank you all so much for your interest
and especially for the commitment you

00:04:41.680 --> 00:04:43.520
show to your students every day.

00:04:43.560 --> 00:04:45.040
So thank you all very much.

00:04:45.040 --> 00:04:47.520
Congratulations and I hope to see you in
many more courses.

00:04:54.480 --> 00:04:58.480
We at the School of Computer Science are
preparing a project called

00:04:58.480 --> 00:05:02.480
Impacta URJC, in which we want to link
students from 4th grade of

00:05:02.480 --> 00:05:06.200
Primary Education to 4th year of
Secondary Education in

00:05:06.200 --> 00:05:09.200
technological projects that have a social
impact.

00:05:09.760 --> 00:05:14.160
Teams will be required to have girls
among their members.

00:05:14.160 --> 00:05:17.880
No, it's not closed to just girls, there
have to be girls present.

00:05:18.320 --> 00:05:21.640
We believe that linking technology with
more social

00:05:21.640 --> 00:05:24.640
action can attract girls to the
technological field.

00:05:24.640 --> 00:05:28.640
So, well, we're also going to add
something else, which is that

00:05:28.640 --> 00:05:32.640
each team will have a mentor, like,
sorry, they will have a

00:05:32.640 --> 00:05:36.640
research professor from our school as a
mentor, so that she can

00:05:36.640 --> 00:05:39.840
also serve as a role model for these
potential

00:05:39.840 --> 00:05:42.840
future students and help to reduce gender
bias.

00:05:42.840 --> 00:05:43.320
Nothing else.

00:05:43.320 --> 00:05:45.960
Congratulations on your participation in
this course and congratulations to the

00:05:45.960 --> 00:05:47.320
organization for

00:05:47.320 --> 00:05:49.960
the great work they have done and to the
Community of Madrid for collaborating on

00:05:49.960 --> 00:05:51.320
this initiative.

00:05:51.320 --> 00:05:51.840
Thank you so much.

00:05:57.280 --> 00:06:00.280
Very well, thank you very much for your
words.

00:06:00.480 --> 00:06:04.480
And well, to use the cooking analogy,
it's my second

00:06:04.480 --> 00:06:07.640
passion after artificial intelligence.

00:06:07.680 --> 00:06:11.680
Today we have a tasting menu with six
courses,

00:06:11.680 --> 00:06:15.320
so let's go, they are very diverse
dishes.

00:06:15.320 --> 00:06:17.920
It is important to know that there are
more technological dishes, more

00:06:17.920 --> 00:06:20.560
educational dishes, and some teaching
experiences.

00:06:20.840 --> 00:06:23.240
Well, a little bit of everything, right?

00:06:23.240 --> 00:06:27.240
Perhaps not all the talks will be for
everyone, in the sense

00:06:27.240 --> 00:06:31.240
that what also happens when one goes to a
tasting menu,

00:06:31.240 --> 00:06:35.240
sometimes I like one dish more than
another, we are going to

00:06:35.240 --> 00:06:39.240
start with a very technological dish and
the first talk will

00:06:39.240 --> 00:06:43.240
be given by Jesús María González
Barahona, from the Rey Juan

00:06:43.240 --> 00:06:47.200
Carlos University, and he will tell us
how to have generative models on your

00:06:47.200 --> 00:06:49.200
computer.

00:06:49.560 --> 00:06:52.200
Note that whenever we have been talking
about generative models up to now, we

00:06:52.200 --> 00:06:53.560
have

00:06:53.560 --> 00:06:56.880
been talking about using generative
models that were in the cloud,

00:06:56.880 --> 00:06:59.880
which were services offered by the big
technology companies.

00:06:59.880 --> 00:07:03.400
It usually works, and Jesus is going to
tell

00:07:03.400 --> 00:07:06.400
us how to do that on our own computers.

00:07:06.680 --> 00:07:10.680
So let's take advantage of the welcoming
applause for Jesus also for the

00:07:10.680 --> 00:07:14.440
farewell applause here at the table, for
the welcome they have given us.

00:07:14.480 --> 00:07:17.200
Thank you all so much. No!

00:07:17.200 --> 00:07:22.600
Director, Vice-Rector, Director, and I
now give the floor to Jesus.

00:07:22.600 --> 00:07:25.600
Thank you all very much.

00:07:32.280 --> 00:07:34.920
First of all, I have no choice but to
thank you for inviting me back here

00:07:34.920 --> 00:07:36.280
again,

00:07:36.280 --> 00:07:40.280
Estefanía, Gregorio and everyone else
involved in the organization.

00:07:40.280 --> 00:07:41.840
Thank you so much.

00:07:41.840 --> 00:07:45.080
Which also gives me the opportunity to
keep updating the talk, as you can

00:07:45.080 --> 00:07:46.720
imagine.

00:07:46.720 --> 00:07:50.280
At the speeds we're going, it's needed
not every year, but every month.

00:07:50.680 --> 00:07:54.680
And in fact, this is a slightly shorter
version of a longer

00:07:54.680 --> 00:07:58.400
version in English, which is the one I've
been using

00:07:58.400 --> 00:08:01.400
lately to talk a little about this type
of model.

00:08:01.720 --> 00:08:05.720
I like to put generative artificial
intelligence running on your own

00:08:05.720 --> 00:08:08.400
infrastructure because it's not something
we necessarily want to run in our own

00:08:08.400 --> 00:08:09.760
home.

00:08:09.760 --> 00:08:12.240
Maybe so, but it's also something that
can be moved.

00:08:12.240 --> 00:08:14.960
For example, a company within the company
itself or in a public administration, as

00:08:14.960 --> 00:08:17.680
is ultimately the case with public
education.

00:08:17.920 --> 00:08:20.560
It could be something that the Ministry
will offer us, for example, or in our

00:08:20.560 --> 00:08:21.920
case, the

00:08:21.920 --> 00:08:25.360
university could be infrastructure
deployed within the university itself.

00:08:25.360 --> 00:08:30.120
The key is that you control the
infrastructure and how you control it.

00:08:30.120 --> 00:08:31.440
Are you sure about what you're doing?

00:08:31.440 --> 00:08:32.800
Think for a moment about what this
implies.

00:08:32.800 --> 00:08:35.240
I'm sure you've thought about it many
times, but let's begin.

00:08:35.240 --> 00:08:38.240
This implies using any artificial
intelligence tool today.

00:08:38.560 --> 00:08:42.400
So you enter, for example, Chile GPT and
you start writing things from that

00:08:42.400 --> 00:08:44.320
moment, you start pressing a key.

00:08:44.320 --> 00:08:47.560
You can rest assured that everything
you're going through there is being

00:08:47.560 --> 00:08:49.200
recorded and stored somewhere.

00:08:49.200 --> 00:08:53.200
And since you normally had to
authenticate first, it's being

00:08:53.200 --> 00:08:57.320
stored and saved along with your data, so
it's perfectly known.

00:08:57.360 --> 00:09:00.000
And now is there someone somewhere, I
don't know, in California, for example,

00:09:00.000 --> 00:09:01.360
who, if you want to know what

00:09:01.360 --> 00:09:04.000
you asked yesterday, what you asked a
month ago, what you asked two months ago,

00:09:04.000 --> 00:09:05.320
what the AI ​​answered you?

00:09:05.320 --> 00:09:06.840
What did you do from there?

00:09:06.840 --> 00:09:09.840
Well, you can know that with complete
peace of mind, which in itself isn't

00:09:09.840 --> 00:09:11.360
necessarily a bad thing.

00:09:11.360 --> 00:09:14.920
In fact, it helps improve systems and
other things, but it's completely out of

00:09:14.920 --> 00:09:16.720
your control.

00:09:16.720 --> 00:09:20.320
And there are two things that are
fundamentally out of your control.

00:09:20.360 --> 00:09:24.360
The first is what happens to your data,
and that's important, but the

00:09:24.360 --> 00:09:28.360
second is perhaps even more important,
and that's how you're actually

00:09:28.360 --> 00:09:32.720
interacting with artificial intelligence,
with the model behind it all.

00:09:32.720 --> 00:09:36.720
Because between you and the artificial
intelligence model there are

00:09:36.720 --> 00:09:40.720
many layers of software that can do
things like change the answer it gives

00:09:40.720 --> 00:09:44.720
you, not give you certain answers, or
even give you different answers

00:09:44.720 --> 00:09:47.800
because it is in the interest of the
provider that does it.

00:09:48.120 --> 00:09:52.120
Furthermore, it should be noted that in
our case, the supplier is

00:09:52.120 --> 00:09:56.120
most likely subject to different
regulations than those we have in

00:09:56.120 --> 00:10:00.120
Spain or Europe, which means that they
may be applying California

00:10:00.120 --> 00:10:03.480
laws to us, and that may not seem
beneficial to us, since

00:10:03.480 --> 00:10:06.480
we did not participate in the enactment
of those laws.

00:10:06.480 --> 00:10:09.440
And there are things that, for example,
are perfectly legal in Spain, but we

00:10:09.440 --> 00:10:10.960
can't do them with AI for that reason.

00:10:10.960 --> 00:10:13.600
Conversely, if you put all of this
together, what happens is that we are in

00:10:13.600 --> 00:10:14.960
a

00:10:14.960 --> 00:10:18.960
situation where we have very little
agency, very little control over how we

00:10:18.960 --> 00:10:22.280
handle artificial intelligence in
general, and as has already been said, I

00:10:22.280 --> 00:10:23.960
think we are all aware of it.

00:10:23.960 --> 00:10:27.520
The use of artificial intelligence will
become increasingly important, not only

00:10:27.520 --> 00:10:29.320
in education, but in all facets of life.

00:10:29.320 --> 00:10:33.320
So, having the possibility of maximizing
that control we have over

00:10:33.320 --> 00:10:36.560
it, at least a priori, seems like an
interesting option.

00:10:37.080 --> 00:10:40.680
And this is an option that allows us to
not have absolute control.

00:10:40.720 --> 00:10:44.200
I will explain why later, but in some
cases, it does require a higher level of

00:10:44.200 --> 00:10:45.960
control than usual.

00:10:45.960 --> 00:10:48.520
Furthermore, it may make sense from an
economic point of view, as you know that

00:10:48.520 --> 00:10:49.840
all this is not free.

00:10:49.840 --> 00:10:52.400
The fact that we can now access some
services for free does not mean that they

00:10:52.400 --> 00:10:53.680
are free.

00:10:53.680 --> 00:10:56.800
We are now paying in other ways, but that
doesn't mean they will always be free or

00:10:56.800 --> 00:10:58.400
at least at that price.

00:10:58.760 --> 00:11:01.400
So, from an economic point of view, in
many cases it might also make sense, but

00:11:01.400 --> 00:11:02.760
today I'm

00:11:02.760 --> 00:11:05.400
not going to focus on the economic
aspect, I'm going to focus on other

00:11:05.400 --> 00:11:06.760
aspects and above

00:11:06.760 --> 00:11:09.360
all I'm going to try to explain it in a
somewhat practical way, not so that you

00:11:09.360 --> 00:11:10.680
can set

00:11:10.680 --> 00:11:13.680
this up in your house, because we
wouldn't have time, would we?

00:11:13.720 --> 00:11:17.720
You'll see later with some of the
technologies that will

00:11:17.720 --> 00:11:21.720
be discussed, but above all so that if
possible you leave

00:11:21.720 --> 00:11:25.720
with a little concern that maybe I should
care a little

00:11:25.720 --> 00:11:29.920
about having control over the AI ​​model
in this version.

00:11:29.920 --> 00:11:33.480
As I was saying, it's a Spanish version
translated from English with one of

00:11:33.480 --> 00:11:36.480
those models that we've installed in my
house in particular.

00:11:36.480 --> 00:11:40.480
I haven't installed this one, I haven't
used it with my home

00:11:40.480 --> 00:11:44.480
hardware, let's say, but later I'll
explain one of the possibilities

00:11:44.480 --> 00:11:48.480
we have of using these models without
having to, as I would say,

00:11:48.480 --> 00:11:52.480
without having to use one of the usual
places where you work in a

00:11:52.480 --> 00:11:56.760
place that gives us more direct access to
the infrastructure.

00:11:56.960 --> 00:11:59.760
That also means the translation has rats.

00:11:59.760 --> 00:12:02.400
It's been checked a little, but it has
some errors, especially in the

00:12:02.400 --> 00:12:03.760
translation of English

00:12:03.760 --> 00:12:06.880
words, which sometimes makes it a bit
strange to translate terms that we use

00:12:06.880 --> 00:12:08.480
differently in Spanish.

00:12:08.480 --> 00:12:13.280
In any case, let's first review what a
generative model has.

00:12:13.280 --> 00:12:17.280
Well, naturally I'm referring throughout
the presentation to AI models,

00:12:17.280 --> 00:12:20.720
let's say, of the latest generation, so
to speak, which is what we normally call

00:12:20.720 --> 00:12:22.480
generative models.

00:12:22.520 --> 00:12:24.800
And I'm going to give a very quick
summary.

00:12:24.800 --> 00:12:28.520
Most of you know that a model is
fundamentally numbers in a neural

00:12:28.520 --> 00:12:30.400
network.

00:12:30.400 --> 00:12:34.400
To put it very simply, each neuron
communicates with the next ones, with the

00:12:34.400 --> 00:12:38.400
parameters, and depending on those
parameters it does some things or others.

00:12:38.440 --> 00:12:40.440
Those parameters are nothing more than
numbers.

00:12:40.440 --> 00:12:43.080
So, if someone gives you a neural
network, if someone gives you a

00:12:43.080 --> 00:12:44.440
generative

00:12:44.440 --> 00:12:47.800
language model, what they give you is a
bunch of numbers.

00:12:47.840 --> 00:12:51.000
And when I say a lot, I usually mean
thousands and millions of numbers.

00:12:52.240 --> 00:12:53.920
But that's all there is to it.

00:12:53.920 --> 00:12:57.920
That, plus information on how to turn
those numbers into something

00:12:57.920 --> 00:13:02.960
that works, which is basically the
software that animates those numbers.

00:13:03.280 --> 00:13:05.680
Therefore, we have what we call the
architecture, which is the way neurons

00:13:05.680 --> 00:13:06.920
are arranged.

00:13:06.920 --> 00:13:10.360
To put it very simply, we also have the
pesos, the numbers in English and in

00:13:10.360 --> 00:13:12.080
Spanish.

00:13:12.080 --> 00:13:15.600
Sometimes we call them weights, sometimes
high, other names, and the software that

00:13:15.600 --> 00:13:17.360
makes it all work.

00:13:17.360 --> 00:13:21.320
The software is usually free software, so
I can download it safely from anywhere.

00:13:21.320 --> 00:13:24.400
Almost all manufacturers, the big
manufacturers, are currently using free

00:13:24.400 --> 00:13:25.960
programs to do it.

00:13:25.960 --> 00:13:28.000
Therefore, that part is not very
complicated.

00:13:28.000 --> 00:13:30.640
The architecture is usually published;
it's part of the technical information,

00:13:30.640 --> 00:13:32.000
and

00:13:32.000 --> 00:13:34.640
they give it to you for various reasons.
The key is whether or not they give me

00:13:34.640 --> 00:13:36.000
the weights.

00:13:36.000 --> 00:13:38.640
If they give me the weights, and I have
the right software and I know the

00:13:38.640 --> 00:13:40.000
architecture,

00:13:40.000 --> 00:13:43.440
I can install it on my own
infrastructure. Sometimes they give me

00:13:43.440 --> 00:13:45.200
the weights, and sometimes they don't.

00:13:45.240 --> 00:13:48.960
But if there are very popular models that
you can no longer access, we still have a

00:13:48.960 --> 00:13:50.840
little bit of a return.

00:13:52.040 --> 00:13:53.720
Okay, let's see, they can fix it.

00:13:53.720 --> 00:13:56.360
Meanwhile, he said that sometimes you can
only use it on the provider's site, let's

00:13:56.360 --> 00:13:57.720
say

00:13:57.720 --> 00:14:00.360
the one that's giving you the service,
but sometimes they also give you the

00:14:00.360 --> 00:14:01.720
possibility,

00:14:01.720 --> 00:14:05.080
if you want, to download those files and
install them on your own machine, on your

00:14:05.080 --> 00:14:06.800
own infrastructure.

00:14:06.800 --> 00:14:09.640
Those are the models I'm going to refer
to.

00:14:09.640 --> 00:14:13.640
There are other things that I won't go
into much detail about today so that the

00:14:13.640 --> 00:14:16.520
models are truly free, reproducible, etc.
In other words, I have total control over

00:14:16.520 --> 00:14:17.960
them.

00:14:18.000 --> 00:14:22.000
I also need other things like training
data, which are essential,

00:14:22.000 --> 00:14:26.000
for example, to detect biases, but also,
for example, to detect

00:14:26.000 --> 00:14:29.520
backdoors or to detect certain
intentional intentions.

00:14:29.520 --> 00:14:30.640
During training.

00:14:30.640 --> 00:14:34.760
I need the full description of the
training process so I can replicate it.

00:14:34.760 --> 00:14:37.400
Then I need the software that allows
training and comparison, because the

00:14:37.400 --> 00:14:38.760
models

00:14:38.760 --> 00:14:42.760
are, let's say, trained in small steps,
and each time there is a new step it is

00:14:42.760 --> 00:14:46.680
compared with the previous one to see if
we have improved or not. I need all of

00:14:46.680 --> 00:14:49.680
that, and therefore I also need all those
intermediate models.

00:14:49.720 --> 00:14:51.760
And I need documentation that explains
all of that.

00:14:51.760 --> 00:14:55.760
When we have all that bottom part, we can
say that we have virtually total

00:14:55.760 --> 00:14:59.440
control over the model, because I could
reproduce that unique model; all I need

00:14:59.440 --> 00:15:01.320
is the right hardware.

00:15:01.360 --> 00:15:04.400
You've probably heard that the right
hardware is expensive, and only big

00:15:04.400 --> 00:15:05.960
companies make it.

00:15:05.960 --> 00:15:09.960
It is true that a lot of hardware is
needed, but it is also true that, for

00:15:09.960 --> 00:15:13.960
example, in Spain or in the European
Union we have more than enough hardware

00:15:13.960 --> 00:15:17.440
to do that kind of thing, paid for in
many cases with public money.

00:15:17.440 --> 00:15:20.000
Whether we want to dedicate it to these
things or whether it's profitable, given

00:15:20.000 --> 00:15:21.320
that

00:15:21.320 --> 00:15:24.320
other people have already done it in
companies, well, it could be done.

00:15:25.040 --> 00:15:29.040
The calculations can be done, and you can
see how if we eventually wanted to have

00:15:29.040 --> 00:15:33.040
a model similar to the big models and we
had been given all that data, we could

00:15:33.040 --> 00:15:35.680
reproduce it, and what is very
interesting, reproduce it with or without

00:15:35.680 --> 00:15:37.040
the

00:15:37.040 --> 00:15:39.680
biases we wanted, and reproduce it
understanding exactly how the model is

00:15:39.680 --> 00:15:41.040
going

00:15:41.040 --> 00:15:45.040
to work, to the point that we understand
it today, although you know that we still

00:15:45.040 --> 00:15:47.680
don't fully understand how these models
work, but we do know that there is a

00:15:47.680 --> 00:15:49.040
certain

00:15:49.040 --> 00:15:52.200
relationship, for example, between the
learning set, the data

00:15:52.200 --> 00:15:55.200
used to train them, and the results that
the model gives.

00:15:55.200 --> 00:15:57.840
We're not going to focus on the ones at
the top, where we mainly have weights,

00:15:57.840 --> 00:15:59.200
although

00:15:59.200 --> 00:16:02.960
I'll talk a little bit about the others
later, because with those at the top my

00:16:02.960 --> 00:16:04.840
control situation improves a lot.

00:16:04.840 --> 00:16:08.280
Even though I can't do everything with
them or know everything about them.

00:16:08.560 --> 00:16:11.760
So we have a wide range of models that
meet these conditions in one way or

00:16:11.760 --> 00:16:13.360
another.

00:16:13.960 --> 00:16:17.080
You know the first ones at the top
because they are, as we usually use

00:16:17.080 --> 00:16:18.680
models, not the majority of people.

00:16:18.680 --> 00:16:21.920
We have an application, usually a web
application, and the model is behind that

00:16:21.920 --> 00:16:23.560
application, so to speak.

00:16:23.560 --> 00:16:25.080
How do we use the web application?

00:16:25.080 --> 00:16:27.720
We don't know exactly what the model
does.

00:16:27.720 --> 00:16:31.720
Then we have the models that are directly
or almost directly accessible. We also

00:16:31.720 --> 00:16:34.720
have some kind of application, but it's a
very lightweight

00:16:34.720 --> 00:16:37.720
projection where what I write practically
goes into the model.

00:16:37.720 --> 00:16:42.920
They're usually GPT-style things, where I
practically talk to the model.

00:16:42.920 --> 00:16:45.560
The ones you have above refer more to
more elaborate things like Notebook, LM.

00:16:45.560 --> 00:16:46.920
For

00:16:46.920 --> 00:16:50.920
example, I give it information and it
generates, for example, a podcast and it

00:16:50.920 --> 00:16:54.880
knows how to do that very well, but there
are other things that I can't do with it.

00:16:55.480 --> 00:16:59.480
Then I have the models with available
weights; these are the ones that will

00:16:59.480 --> 00:17:03.480
start to interest me from then on. These
are the ones I've talked about before,

00:17:03.480 --> 00:17:06.120
where the manufacturer, the one who
trained the model, has given me the

00:17:06.120 --> 00:17:07.480
weights.

00:17:07.480 --> 00:17:11.480
And then I have several categories, which
are where the information I had

00:17:11.480 --> 00:17:15.480
before, where I was saying that I can
also reproduce what I have, the set of

00:17:15.480 --> 00:17:19.480
elements, etc. Well, depending on how
many of those characteristics I have,

00:17:19.480 --> 00:17:21.960
the others below in the complete
presentation have a description of all of

00:17:21.960 --> 00:17:23.240
them.

00:17:23.280 --> 00:17:24.320
But I'm not going in today.

00:17:24.320 --> 00:17:26.960
But I would like you to understand that
what we are going to talk about when we

00:17:26.960 --> 00:17:28.320
say to

00:17:28.320 --> 00:17:30.960
arrange the weights, as I said before, is
only a certain improvement in control

00:17:30.960 --> 00:17:32.320
that

00:17:32.320 --> 00:17:36.240
may be very interesting to me but that is
not absolute, that there are still more

00:17:36.240 --> 00:17:38.240
things we can achieve and that suits me.

00:17:38.240 --> 00:17:42.240
This would be the complete scheme in
English, where what we have in the

00:17:42.240 --> 00:17:46.240
middle, the greenest part, is what we
would love to have. That is, models

00:17:46.240 --> 00:17:50.240
that we control almost totally where we
know how the model is trained, we

00:17:50.240 --> 00:17:54.240
know what has happened to it, we can
reproduce it from scratch if we want,

00:17:54.240 --> 00:17:58.240
etc. And at the top end we have the
models that we use behind applications

00:17:58.240 --> 00:18:01.960
where we can only interact with them in
very limited and very controlled ways.

00:18:04.320 --> 00:18:08.040
In this context we can see that we have
three fundamental things that

00:18:08.040 --> 00:18:11.040
matter to us in determining the openness
of a model.

00:18:11.040 --> 00:18:14.200
On the one hand we can say that the model
itself is open.

00:18:14.200 --> 00:18:17.640
To sum it up, just give me the weights
and the information I need to be able to

00:18:17.640 --> 00:18:19.360
handle it.

00:18:19.920 --> 00:18:21.640
But we also have the software.

00:18:21.640 --> 00:18:24.840
As we've said, the software itself can be
open source software, right?

00:18:24.840 --> 00:18:27.000
And I'm referring to the software for
making inferences, that is, for

00:18:27.000 --> 00:18:28.120
interacting with

00:18:28.120 --> 00:18:31.120
the model when it is already trained, but
also for training it, for example.

00:18:31.520 --> 00:18:35.520
And finally, we have what we might call
open science characteristics, which are

00:18:35.520 --> 00:18:38.520
things like everything being reproducible
from scratch.

00:18:39.960 --> 00:18:44.440
Each of those things you have there can
be studied separately and in each model.

00:18:44.440 --> 00:18:48.440
If I want to know exactly what I can and
cannot do with the model, I need to know

00:18:48.440 --> 00:18:52.040
for each of them what characteristics it
has, what they give me in return, and

00:18:52.040 --> 00:18:53.880
under what license.

00:18:53.880 --> 00:18:57.880
Because although the legal status of the
models is still under much legal

00:18:57.880 --> 00:19:01.920
discussion, for now licenses very similar
to those of software are being used.

00:19:01.920 --> 00:19:04.560
Therefore, they can give me the model but
without permission, for example for

00:19:04.560 --> 00:19:05.920
commercial use or without

00:19:05.920 --> 00:19:08.880
permission for I don't know what or for I
don't know how many, for example, without

00:19:08.880 --> 00:19:11.880
permission to retrain, which is something
that may interest me a lot.

00:19:11.880 --> 00:19:14.520
So sometimes I have the technical ability
to do something, but I don't have the

00:19:14.520 --> 00:19:15.880
legal ability

00:19:15.880 --> 00:19:19.000
because the manufacturer who made the
model, the one who trained it, doesn't

00:19:19.000 --> 00:19:20.600
give me that possibility.

00:19:20.600 --> 00:19:23.480
In any case, as I said, doing a detailed
analysis involves looking at all those

00:19:23.480 --> 00:19:24.960
things.

00:19:25.000 --> 00:19:29.000
I just want you to understand that
talking about open models is a huge

00:19:29.000 --> 00:19:32.480
oversimplification, and I'm going to do
that from now on.

00:19:32.520 --> 00:19:33.480
But. But it's very big.

00:19:33.480 --> 00:19:37.480
There are many details I need to look at
to see exactly what it means for the

00:19:37.480 --> 00:19:40.120
model to be open, because for each of
those things we can apply the four

00:19:40.120 --> 00:19:41.480
freedoms

00:19:41.480 --> 00:19:45.480
of free software that you may know, which
were defined by Richard Stallman

00:19:45.480 --> 00:19:49.480
back in the 80s and which are the four
things that are looked at to know if a

00:19:49.480 --> 00:19:53.480
program is free: that I can use it as I
want, that I can study it as I want, that

00:19:53.480 --> 00:19:57.400
I can modify it, if I am interested in
modifying it, and that I can redistribute

00:19:57.400 --> 00:19:59.400
it, give it to others, etc., etc.

00:19:59.440 --> 00:20:02.800
I can apply the same thing to models, but
I can also apply it to all the things

00:20:02.800 --> 00:20:04.480
we've seen before about models.

00:20:04.520 --> 00:20:07.160
I can apply it to the weights, I can
apply it to the inference software, it

00:20:07.160 --> 00:20:08.520
can

00:20:08.520 --> 00:20:12.360
be applied to training software, I can
only apply the dataset, etc., etc.

00:20:12.680 --> 00:20:16.400
So when I want to know if a model is,
let's say, completely free, it would have

00:20:16.400 --> 00:20:18.280
to meet all those conditions.

00:20:18.280 --> 00:20:19.480
For all of the above.

00:20:19.480 --> 00:20:22.840
There are some of those models, but they
are very few for primarily

00:20:22.840 --> 00:20:25.840
practical reasons, as I will also discuss
a little later.

00:20:27.840 --> 00:20:30.480
All of this ultimately boils down to a
few things that are what I might be more

00:20:30.480 --> 00:20:31.840
or

00:20:31.840 --> 00:20:35.600
less interested in knowing, such as
whether or not I can have a model, for

00:20:35.600 --> 00:20:37.520
example, what level of access I have.

00:20:37.600 --> 00:20:39.320
I can run the model however I want.

00:20:39.320 --> 00:20:41.200
That's technically called making
inferences.

00:20:41.200 --> 00:20:43.840
I can ask any question, so to speak, the
model, or there will be certain questions

00:20:43.840 --> 00:20:45.200
that I

00:20:45.200 --> 00:20:47.920
won't be able to ask or that the model
won't answer, but not because the model

00:20:47.920 --> 00:20:50.680
doesn't know, but because they are
restricted.

00:20:50.680 --> 00:20:54.680
Because in my access to the model, those
types of questions aren't allowed,

00:20:54.680 --> 00:20:58.680
and there's something in between that
intercepts the response and removes

00:20:58.680 --> 00:21:02.320
parts of it so I don't see them. Or, as I
said before, it changes them.

00:21:02.320 --> 00:21:04.120
Everyone here thinks about sex and
weapons.

00:21:04.120 --> 00:21:06.600
I don't know what to think, about more
mundane things, like you asking "Where do

00:21:06.600 --> 00:21:07.880
I give?"

00:21:08.120 --> 00:21:11.240
This summer I'm going to Valencia on
vacation where I can go for a walk.

00:21:11.440 --> 00:21:14.360
Imagine that the model captures the
response and where it says "Well, go for

00:21:14.360 --> 00:21:15.840
a walk on the beach."

00:21:15.840 --> 00:21:16.680
I don't know what he's telling me.

00:21:16.680 --> 00:21:20.680
Go get a coffee at Starbucks because the
model is programmed that way, because

00:21:20.680 --> 00:21:24.680
it's software that interacts with the
model, and every time it finds a pattern

00:21:24.680 --> 00:21:28.680
like that, it changes it for someone who
has sponsored it, for example.

00:21:29.040 --> 00:21:32.160
But these are things we're not yet seeing
from a sponsorship standpoint, but I can

00:21:32.160 --> 00:21:33.720
assure you we will see them very soon.

00:21:33.880 --> 00:21:36.240
And above all, it's very difficult to
know when they are happening, when they

00:21:36.240 --> 00:21:37.440
are not happening.

00:21:37.640 --> 00:21:41.040
So that level of access is, let's say,
the first thing I'd like to know.

00:21:41.040 --> 00:21:43.680
I have real access to the model and the
limitations I'm seeing are those of the

00:21:43.680 --> 00:21:45.040
model

00:21:45.040 --> 00:21:48.480
itself, not those of the software or
things that have been put in front of me.

00:21:48.840 --> 00:21:52.280
Second control: control has to do with
being able

00:21:52.280 --> 00:21:55.280
to modify the way it works technically.

00:21:55.360 --> 00:21:56.920
There are many ways to do it.

00:21:56.920 --> 00:22:00.360
One of the most common is Hay Tunes,
which has been translated for me here in

00:22:00.360 --> 00:22:02.120
a very, very funny way.

00:22:02.120 --> 00:22:05.240
Finally, "tuning"—which I don't really
know what "tuning" means—is like refining

00:22:05.240 --> 00:22:06.840
the model.

00:22:06.840 --> 00:22:10.280
This means taking the model as it is and
partially retraining it to perform

00:22:10.280 --> 00:22:12.040
specific tasks.

00:22:12.040 --> 00:22:15.080
It's a very interesting thing because it
allows a generalist model to be very good

00:22:15.080 --> 00:22:16.640
at certain things.

00:22:16.640 --> 00:22:17.880
I give him some documentation.

00:22:17.880 --> 00:22:21.880
For example, imagine a medical exam for
something or other,

00:22:21.880 --> 00:22:25.880
and the model learns that documentation
specifically and

00:22:25.880 --> 00:22:31.080
is able to answer questions related to
that documentation much better.

00:22:31.120 --> 00:22:34.600
And is that something we can all do? Can
we give him our course

00:22:34.600 --> 00:22:37.600
notes, for example? Can I do that kind of
thing, or not?

00:22:37.600 --> 00:22:41.600
I can evolve the model in ways that
interest me or not; this is fundamental

00:22:41.600 --> 00:22:44.600
to being able to use the model
realistically and, above all, more

00:22:44.600 --> 00:22:46.120
reliably.

00:22:46.920 --> 00:22:48.320
What control do we have afterwards?

00:22:48.320 --> 00:22:51.360
I've already talked about the data, and
before, I can control what I pass to the

00:22:51.360 --> 00:22:52.920
model.

00:22:52.920 --> 00:22:55.600
Can I monitor the results? Yes, if they
are being saved.

00:22:55.600 --> 00:22:58.760
I don't know if those are the ones I'm
actually putting in.

00:22:59.240 --> 00:23:00.400
Autonomy.

00:23:00.400 --> 00:23:02.080
How dependent are we?

00:23:02.080 --> 00:23:04.720
For example, if you start using a model
in class, you will be able to use it

00:23:04.720 --> 00:23:06.080
throughout the

00:23:06.080 --> 00:23:10.080
academic year, or perhaps halfway through
the year they will change it because the

00:23:10.080 --> 00:23:12.720
manufacturer is interested and what was
previously free now becomes a paid

00:23:12.720 --> 00:23:14.080
service, or

00:23:14.080 --> 00:23:16.680
they simply remove access to the model
because it is no longer commercially

00:23:16.680 --> 00:23:18.000
viable for them.

00:23:19.440 --> 00:23:22.080
If I don't have a minimum level of
independence that allows me to decide

00:23:22.080 --> 00:23:23.440
when

00:23:23.440 --> 00:23:27.440
I want to use the model, but instead I
depend on the manufacturer, my autonomy

00:23:27.440 --> 00:23:31.520
is very limited, and mind you, autonomy
is also linked to the economy.

00:23:31.560 --> 00:23:34.200
The autonomy might be related to the fact
that it's free for me now, but halfway

00:23:34.200 --> 00:23:35.560
through the course,

00:23:35.560 --> 00:23:38.240
they charge me a certain amount because
the manufacturer's policy has changed,

00:23:38.240 --> 00:23:39.600
for example.

00:23:39.600 --> 00:23:42.240
So my autonomy is very limited,
especially if I don't have those

00:23:42.240 --> 00:23:43.600
resources

00:23:43.600 --> 00:23:46.680
to spend on the model or if my students
don't have them.

00:23:47.040 --> 00:23:49.520
And finally, trust. I've missed a line
here.

00:23:49.520 --> 00:23:50.360
I'm sorry.

00:23:50.360 --> 00:23:54.160
Basically, I can make sure that the model
does what I think it does.

00:23:54.600 --> 00:23:58.160
Or as I said before, that they are
changing the behavior of the model in

00:23:58.160 --> 00:23:59.960
some way.

00:24:01.160 --> 00:24:02.480
Why does all this matter?

00:24:02.480 --> 00:24:03.800
Well, you can imagine.

00:24:03.800 --> 00:24:07.800
Access to the model matters to me because
it enables or prevents many use

00:24:07.800 --> 00:24:11.080
cases, or there are things I won't be
able to do if I don't have the access I

00:24:11.080 --> 00:24:12.760
want. The model.

00:24:12.760 --> 00:24:15.520
But it also has a lot to do with the
possibilities for innovation and the

00:24:15.520 --> 00:24:18.280
possibilities for integration with other
tools.

00:24:18.280 --> 00:24:20.920
In other words, realize that it's not
just a matter of whether or not I'm

00:24:20.920 --> 00:24:22.280
deceived by the

00:24:22.280 --> 00:24:24.840
answers, it's a matter of whether or not
I can develop applications from within a

00:24:24.840 --> 00:24:26.160
company.

00:24:26.160 --> 00:24:27.720
For example.

00:24:27.720 --> 00:24:31.720
Control of the model also has a lot to do
with what use cases it can utilize,

00:24:31.720 --> 00:24:35.720
whether it can be used, for example, by
companies or public institutions in

00:24:35.720 --> 00:24:38.720
certain cases, and also with innovation
and integration.

00:24:38.760 --> 00:24:42.760
Data control obviously has a huge impact
on privacy, but also

00:24:42.760 --> 00:24:46.360
on our dependence on who provides us with
those models.

00:24:46.880 --> 00:24:50.600
Also about this asymmetry of knowledge,
that they are going to have a

00:24:50.600 --> 00:24:53.600
lot of knowledge about us, much more than
they have now.

00:24:53.720 --> 00:24:56.720
And that, depending on the circumstances,
might not interest us.

00:24:56.720 --> 00:25:00.240
And again, remember that this affects us
not only as individuals, but also, for

00:25:00.240 --> 00:25:02.040
example, as organizations.

00:25:02.120 --> 00:25:04.760
Is there a place where they will know
everything my company or organization is

00:25:04.760 --> 00:25:06.120
doing

00:25:06.120 --> 00:25:08.760
because it's receiving all the requests
from my employees and often directly

00:25:08.760 --> 00:25:10.120
linked

00:25:10.120 --> 00:25:12.920
to the company name, for example, which
is the one who pays for that, and in

00:25:12.920 --> 00:25:15.760
certain contexts that can be very
important?

00:25:15.760 --> 00:25:18.000
Imagine the context of defense, for
example.

00:25:18.000 --> 00:25:20.640
Then I'll explain why there are so many
defense companies, because the use of

00:25:20.640 --> 00:25:22.000
these

00:25:22.000 --> 00:25:25.040
models is prohibited, for example,
although people then use ways to try to

00:25:25.040 --> 00:25:26.600
continue using them.

00:25:27.000 --> 00:25:31.000
Well, autonomy has already been mentioned
and it has to do with very basic

00:25:31.000 --> 00:25:33.640
aspects such as competence and
competitiveness and trust, which has to

00:25:33.640 --> 00:25:35.000
do

00:25:35.000 --> 00:25:39.000
not only with everything we have talked
about, but also with transparency,

00:25:39.000 --> 00:25:43.520
how much we know about what the model is
doing versus what we believe it is doing.

00:25:45.480 --> 00:25:49.480
Okay, now you have several slides that
I'm going to skip over, and the

00:25:49.480 --> 00:25:53.000
laws afterwards, if you want, are each of
the types of access to the model that

00:25:53.000 --> 00:25:54.760
I've put.

00:25:54.760 --> 00:25:58.720
So what are the implications from these
points of view that I've mentioned?

00:25:58.720 --> 00:26:00.960
I'm going to skip them because we have a
chart down here and you have, by the way,

00:26:00.960 --> 00:26:02.120
a lot of links.

00:26:02.120 --> 00:26:05.920
You can download the presentation later
if you want, and there are

00:26:05.920 --> 00:26:08.920
many links and sites where you can use
all these things.

00:26:09.560 --> 00:26:13.000
But I'm going to skip ahead to a table
that's not here, because I'll arrive on

00:26:13.000 --> 00:26:14.760
time in the end.

00:26:14.760 --> 00:26:18.520
Remember, this was the diagram with all
the models and this is the table I was

00:26:18.520 --> 00:26:20.400
referring to.

00:26:20.400 --> 00:26:24.240
This is a very simplified table regarding
the type of access to the model.

00:26:24.240 --> 00:26:25.520
What can I do then?

00:26:25.520 --> 00:26:29.400
One first thing we can do is, for
example, when I am working with

00:26:29.400 --> 00:26:32.400
her or when I am asking my students to
work with her.

00:26:32.440 --> 00:26:36.440
I can see which of those parts I'm in,
and if I haven't made a mistake

00:26:36.440 --> 00:26:39.640
in the classification, then we
immediately see what characteristics that

00:26:39.640 --> 00:26:41.240
has.

00:26:41.240 --> 00:26:45.240
For example, when we are asking our
students to use, for example,

00:26:45.240 --> 00:26:49.240
an automatic translator based on
generative AI models, then the

00:26:49.240 --> 00:26:52.800
access is to translate as defined by the
app, nothing more.

00:26:52.840 --> 00:26:56.840
And in all other dimensions we have
nothing, we have no control over the

00:26:56.840 --> 00:27:00.680
model, we don't know if it's doing other
things or any other story.

00:27:00.680 --> 00:27:03.400
We have no control over the data, what we
ask you to translate, or what you return

00:27:03.400 --> 00:27:04.760
to us. Translated.

00:27:04.760 --> 00:27:06.160
We have no autonomy.

00:27:06.160 --> 00:27:08.600
If they cut off my access to the model,
there's nothing I can do.

00:27:08.600 --> 00:27:12.600
And finally, I can only be confident or
not that the translation is good

00:27:12.600 --> 00:27:16.600
depending on how I look at the
translation, but it is impossible to be

00:27:16.600 --> 00:27:20.600
sure that specific translations will be
done correctly, because no,

00:27:20.600 --> 00:27:24.600
I have no real information about how the
model works at the opposite

00:27:24.600 --> 00:27:28.600
extreme, which is the one you have at the
very bottom, which would be,

00:27:28.600 --> 00:27:32.600
so to speak, the most open use than what
we call reproducible, because

00:27:32.600 --> 00:27:35.960
it means that I could use it however I
want because I would use it in my

00:27:35.960 --> 00:27:37.680
infrastructure.

00:27:37.680 --> 00:27:40.320
Furthermore, I would control all the
software, I would control the model, I

00:27:40.320 --> 00:27:41.680
knew how the model was trained,

00:27:41.680 --> 00:27:44.320
so I also have complete control of the
model, I also have complete control of

00:27:44.320 --> 00:27:45.680
the data, because I know

00:27:45.680 --> 00:27:48.920
everything that happens to the model and
everything that the model returns to me.

00:27:48.920 --> 00:27:51.200
It's all within my environment, so to
speak.

00:27:51.200 --> 00:27:54.560
Is the autonomy complete because I am the
provider myself, or is it someone I trust

00:27:54.560 --> 00:27:56.240
who is doing it for me?

00:27:56.240 --> 00:27:58.880
The provision of the service and the
confidence that the model does what it

00:27:58.880 --> 00:28:00.240
should do,

00:28:00.240 --> 00:28:03.480
because somehow the model does exactly
what it has been trained to do, which I

00:28:03.480 --> 00:28:05.120
also know it is, is total.

00:28:05.320 --> 00:28:09.320
So of course, keep in mind that all our
models will fall between those two

00:28:09.320 --> 00:28:13.320
extremes, and as I mentioned before, in
the time I have left, I'm going to

00:28:13.320 --> 00:28:16.600
focus on those of you who have models
with lower available weights.

00:28:16.760 --> 00:28:20.760
In other words, let's at least assume
that I can run the model on our own

00:28:20.760 --> 00:28:24.080
infrastructure and realize that this
already gives me a few

00:28:24.080 --> 00:28:27.080
interesting things compared to what we
had at the top.

00:28:27.120 --> 00:28:30.800
I suspect, for example, that access will
be somewhat restricted,

00:28:30.800 --> 00:28:33.800
but I'll already have many interesting
features.

00:28:33.800 --> 00:28:36.480
I don't know everything the model does,
but I know that I put in the software

00:28:36.480 --> 00:28:39.200
that manages the model, for example, so I
can trust it.

00:28:39.200 --> 00:28:42.600
The model control isn't perfect either,
because I don't know exactly if the model

00:28:42.600 --> 00:28:44.320
does what I've been told it does.

00:28:44.320 --> 00:28:47.400
It may be conditioned, it may have
backdoors, but again

00:28:47.400 --> 00:28:50.400
it's on my infrastructure running with my
software.

00:28:50.400 --> 00:28:53.440
So I have considerably more control than
I did before.

00:28:53.680 --> 00:28:57.680
Data control is complete, and that is
fundamental for issues such as privacy,

00:28:57.680 --> 00:29:00.560
and especially for the privacy of our
students, which is particularly

00:29:00.560 --> 00:29:02.040
protected.

00:29:02.040 --> 00:29:05.240
Instead of sending the data to a company
that I have to trust or not to do who

00:29:05.240 --> 00:29:06.880
knows what with it, I send it myself.

00:29:06.880 --> 00:29:09.520
Imagine in Madrid, for example, we send
it to the Ministry, whatever

00:29:09.520 --> 00:29:10.880
infrastructure

00:29:10.880 --> 00:29:14.880
they have there and we directly trust the
Ministry because they have set up

00:29:14.880 --> 00:29:18.880
infrastructure precisely so that this can
be done without any problems of privacy

00:29:18.880 --> 00:29:21.520
and autonomy and trust, without being
perfect, well in terms of autonomy I have

00:29:21.520 --> 00:29:22.880
a

00:29:22.880 --> 00:29:26.400
certain autonomy because I provide the
service and so on and I have the weights,

00:29:26.400 --> 00:29:28.200
so I can do quite a few things.

00:29:28.200 --> 00:29:31.680
I have very little confidence because I
don't know exactly what's in the model.

00:29:31.680 --> 00:29:34.320
Remember that these are models whose
training methods I don't know; they give

00:29:34.320 --> 00:29:35.680
me the weights

00:29:35.680 --> 00:29:38.720
and the model's characteristics, but I
don't know how they were trained.

00:29:38.720 --> 00:29:42.720
Therefore, I can have less confidence
there, and in between we have several

00:29:42.720 --> 00:29:45.600
degrees of open models, some with more
interesting features and others with

00:29:45.600 --> 00:29:47.080
fewer.

00:29:48.560 --> 00:29:50.160
Many problems remain unresolved.

00:29:50.160 --> 00:29:54.360
Anyway, with this I just want to
introduce you a little to the problem.

00:29:54.360 --> 00:29:57.840
For example, we have a very complex
relationship between the training data.

00:29:57.840 --> 00:29:58.600
The model.

00:29:58.600 --> 00:30:02.600
Reproducing it is not trivial for many
reasons, not only because of the cost and

00:30:02.600 --> 00:30:06.240
infrastructure, and in many cases, it is
also not easy to know exactly which

00:30:06.240 --> 00:30:09.240
categories to look at in a model. Many
details and not always.

00:30:09.240 --> 00:30:11.160
It's not always easy to do.

00:30:11.160 --> 00:30:15.120
And perhaps the most important point is
the one at the very bottom.

00:30:15.120 --> 00:30:17.760
If I don't have absolutely all the data
for the model and nobody has reproduced

00:30:17.760 --> 00:30:19.120
it, I

00:30:19.120 --> 00:30:22.040
can believe what the manufacturer tells
me or not, but it's very difficult to

00:30:22.040 --> 00:30:24.960
know that the data they are giving me is
correct.

00:30:26.240 --> 00:30:30.240
And with this we are now going to talk
very quickly about some examples

00:30:30.240 --> 00:30:33.520
and some direct implications of these
models, and these are the ones I can

00:30:33.520 --> 00:30:35.200
install.

00:30:35.200 --> 00:30:37.520
I repeat, in terms of infrastructure at
least they have to give me the money for

00:30:37.520 --> 00:30:38.680
the

00:30:38.680 --> 00:30:41.680
software and, let's say, the recipe, the
architecture to be able to run it.

00:30:43.080 --> 00:30:44.840
So, what's the minimum?

00:30:44.840 --> 00:30:48.840
So we go back again to reviewing the
inference code with all the necessary

00:30:48.840 --> 00:30:52.040
libraries, which fortunately is usually
free software.

00:30:52.080 --> 00:30:55.880
Therefore, most or all of the models I
know of give me the weights, that's not a

00:30:55.880 --> 00:30:57.800
problem.

00:30:58.280 --> 00:31:02.280
Then the weights, including some
parameters, etc., which they also have to

00:31:02.280 --> 00:31:05.200
give me to use the model, basically so
that I can run it with the software I

00:31:05.200 --> 00:31:06.680
wrote earlier.

00:31:06.720 --> 00:31:10.720
So if they give it to me, perfect, I
usually also need at least a technical

00:31:10.720 --> 00:31:14.720
report, at least a summary, where they
explain how I have to do the

00:31:14.720 --> 00:31:18.720
prompts, what kind of capabilities the
model has, things like that.

00:31:18.720 --> 00:31:21.560
For example, whether the model
understands a graph or not, whether the

00:31:21.560 --> 00:31:24.440
model is capable of producing graphs or
only text, things like that.

00:31:24.480 --> 00:31:27.120
And usually there's something called the
model card, which is the technical

00:31:27.120 --> 00:31:28.480
specifications

00:31:28.480 --> 00:31:31.680
of the model, which gives me some more
detailed information that helps me run it

00:31:31.680 --> 00:31:33.280
with confidence.

00:31:33.280 --> 00:31:36.520
Let's say the minimum, I repeat once
again, is a model with

00:31:36.520 --> 00:31:39.520
available weights, just as we classified
it before.

00:31:42.040 --> 00:31:43.800
I'm skipping this because I don't think
it says anything new.

00:31:43.800 --> 00:31:47.800
Let's go back to the previous chart and
look at some examples of where

00:31:47.800 --> 00:31:52.360
these models live. First, the place where
they usually live is called Aging Face.

00:31:52.360 --> 00:31:55.640
For those of you who know GitHub, the
world of software, well, it's a bit like

00:31:55.640 --> 00:31:57.320
the GitHub of models.

00:31:57.320 --> 00:32:02.280
There are literally tens of thousands of
models hosted on face.

00:32:02.320 --> 00:32:04.880
I can download the weights of all of
them.

00:32:04.880 --> 00:32:08.080
Not all of them are models of what they
call foundations,

00:32:08.080 --> 00:32:11.080
models that someone has trained from
scratch and so on.

00:32:11.080 --> 00:32:15.080
Many of them are simply refinements of
other models, adaptations,

00:32:15.080 --> 00:32:18.560
etc. But there are models of all kinds
for all kinds of

00:32:18.560 --> 00:32:21.560
tasks and for all kinds of activities.

00:32:21.640 --> 00:32:24.840
So, if you go to Facebook and look for
"Models,"

00:32:24.840 --> 00:32:27.840
you can find a very complete list of all
of them.

00:32:28.520 --> 00:32:32.080
And the main advantages of these models
have to

00:32:32.080 --> 00:32:35.080
do with the features we have seen before.

00:32:35.080 --> 00:32:37.000
Perhaps the first one is the most
important.

00:32:37.000 --> 00:32:39.600
We know that once I have the model, they
can't take it away from me.

00:32:39.600 --> 00:32:42.000
I will always have it available, and
that's essential.

00:32:42.000 --> 00:32:46.000
If I want to ensure that I can do things
with the model, I can execute them.

00:32:46.000 --> 00:32:48.640
Infrastructure, as I said, I can
integrate them very easily with other

00:32:48.640 --> 00:32:50.000
applications and in

00:32:50.000 --> 00:32:52.640
fact, one of the most remarkable things
that is happening now in this world is

00:32:52.640 --> 00:32:54.000
precisely that

00:32:54.000 --> 00:32:56.640
they are integrating with programs that
do lots of different things, so that even

00:32:56.640 --> 00:32:58.000
if you don't

00:32:58.000 --> 00:33:00.960
know anything about computers, you can
simply be the model, sorry, the program

00:33:00.960 --> 00:33:02.440
that is very fast.

00:33:02.440 --> 00:33:05.120
You download the template, which is
usually transparent because the program

00:33:05.120 --> 00:33:07.800
itself creates it, and you can start
using it directly.

00:33:08.840 --> 00:33:12.840
There is a lot of competition in the
market because in the end any company

00:33:12.840 --> 00:33:16.240
can install them and provide access to
them via the Internet.

00:33:16.280 --> 00:33:20.280
And if you decide, you're the one who
decides whether you want to run them

00:33:20.280 --> 00:33:22.920
either in the cloud on some
infrastructure that someone can see, or

00:33:22.920 --> 00:33:24.280
on

00:33:24.280 --> 00:33:28.280
your computer, or sometimes even on your
mobile phone, depending on the

00:33:28.280 --> 00:33:32.560
model; they also have disadvantages, it
takes a little more work to use them.

00:33:33.360 --> 00:33:36.320
They are usually not as advanced as the
most advanced ones available.

00:33:36.320 --> 00:33:38.880
Right now, people are usually talking
about a delay, like between a year and a

00:33:38.880 --> 00:33:40.200
year and a half.

00:33:40.200 --> 00:33:42.840
If you mean that the open models of
today, well, maybe they are equivalent to

00:33:42.840 --> 00:33:44.200
those of, I

00:33:44.200 --> 00:33:47.120
don't know, January of 25, although other
people say that this has been reduced a

00:33:47.120 --> 00:33:50.040
lot and that now we are between six and
nine months.

00:33:50.040 --> 00:33:54.040
But it depends a lot on the model and the
task you're interested in. And we've

00:33:54.040 --> 00:33:57.440
already seen the advantages of
self-hosting, that is, setting it up on

00:33:57.440 --> 00:33:59.160
your own infrastructure.

00:33:59.160 --> 00:34:02.160
So I'm going to skip those.

00:34:02.560 --> 00:34:05.200
I'm going to focus for a moment on the
disadvantages, which also exist,

00:34:05.200 --> 00:34:06.560
remembering

00:34:06.560 --> 00:34:09.640
that they are running on infrastructure
that I provide.

00:34:09.680 --> 00:34:12.320
So, access isn't going to be free for me;
I'm going to have to buy or pay for the

00:34:12.320 --> 00:34:13.680
infrastructure

00:34:13.680 --> 00:34:16.320
somehow. And I'm wondering how it's
possible that if the infrastructure costs

00:34:16.320 --> 00:34:17.680
so

00:34:17.680 --> 00:34:20.320
much—and it's true that it costs so
much—they're now giving me free access to

00:34:20.320 --> 00:34:21.680
it because

00:34:21.680 --> 00:34:24.720
everyone who uses, for example, that GPT,
knows that within certain limits, you get

00:34:24.720 --> 00:34:26.280
it for free.

00:34:26.280 --> 00:34:28.920
But start thinking right away if it
hasn't already happened to you, or if you

00:34:28.920 --> 00:34:30.280
know

00:34:30.280 --> 00:34:34.280
people who have, that you've already gone
beyond those limits and started buying

00:34:34.280 --> 00:34:37.080
it, because when you start using it, you
use it very little, but over time you use

00:34:37.080 --> 00:34:39.880
it more and you probably need more
things.

00:34:39.880 --> 00:34:43.880
And above all, consider whether prices
will remain that way or if

00:34:43.880 --> 00:34:47.240
there are incentives for prices not to
remain that way.

00:34:47.600 --> 00:34:51.080
In any case, we have to maintain it and
we have to pay for it.

00:34:51.080 --> 00:34:55.080
And besides, the available hardware is
expensive, especially now,

00:34:55.080 --> 00:34:59.080
when there is a great need for many
companies to acquire that hardware,

00:34:59.080 --> 00:35:02.760
so its price is not decreasing at the
same rate as other hardware.

00:35:02.760 --> 00:35:06.760
In IT, you have to take care of
maintenance, even if it's small,

00:35:06.760 --> 00:35:10.680
and above all, you have to make many
decisions to stay up to date.

00:35:10.680 --> 00:35:14.680
This world moves very fast, and the model
that was better a year ago is now one

00:35:14.680 --> 00:35:18.400
of the worst, so you have to be very
aware of how things are evolving.

00:35:18.400 --> 00:35:22.400
In other words, right now this is still a
world where you'll need mostly

00:35:22.400 --> 00:35:26.400
knowledge and a few resources—not a lot
of resources without a lot of

00:35:26.400 --> 00:35:30.400
knowledge, mind you, not special
technical knowledge like knowing a lot

00:35:30.400 --> 00:35:33.280
about computers, but understanding very
well which tools are the best here right

00:35:33.280 --> 00:35:34.760
now.

00:35:34.800 --> 00:35:37.320
And which models are the best, and then
how to use them.

00:35:37.320 --> 00:35:39.960
But understand that there are many people
involved in this who don't have specific

00:35:39.960 --> 00:35:41.320
training

00:35:41.320 --> 00:35:44.040
in computer science; what they do have is
specific training in how these types of

00:35:44.040 --> 00:35:45.440
models work.

00:35:45.520 --> 00:35:49.520
If we have time, I'll give an example
later; we'll get into the

00:35:49.520 --> 00:35:53.240
equipment requirements, as they depend a
lot on what you want to do.

00:35:53.280 --> 00:35:57.280
A high-end computer with a gaming GPU
like the ones kids use

00:35:57.280 --> 00:36:01.080
for gaming is probably more than enough
to run most of the

00:36:01.080 --> 00:36:04.080
games and do some pretty interesting
things.

00:36:04.440 --> 00:36:08.680
Little by little things are changing too,
and less and less hardware is needed.

00:36:08.680 --> 00:36:11.200
There are now models that even run on
mobile phones.

00:36:11.200 --> 00:36:15.080
The thing is, they might not give you the
same capabilities as a model that runs

00:36:15.080 --> 00:36:17.040
the cloud with a bunch of GPUs.

00:36:17.040 --> 00:36:20.920
But in principle, as technology advances.

00:36:20.960 --> 00:36:24.680
We are also getting smaller models that
fit in smaller devices,

00:36:24.680 --> 00:36:27.680
with less specific hardware and are
cheaper.

00:36:31.600 --> 00:36:34.760
I'll skip over a bit of this, I'll just
stick with the last line.

00:36:35.000 --> 00:36:37.640
In the end, you have to do the math if
you want to set it up at home, but

00:36:37.640 --> 00:36:39.000
nowadays

00:36:39.000 --> 00:36:43.000
this is, I would say, especially
designed, as I said before, for those who

00:36:43.000 --> 00:36:47.000
like the subject and turn it into their
hobby and want to do these things, or

00:36:47.000 --> 00:36:51.000
for organizations that see it as a way to
improve either the economy or

00:36:51.000 --> 00:36:54.960
autonomy, or any of these other values
​​we have been talking about before.

00:36:56.080 --> 00:36:59.800
And we have a very interesting option
which

00:36:59.800 --> 00:37:02.800
is to use auto-talk models via APIs.

00:37:03.320 --> 00:37:04.240
What does this mean?

00:37:04.240 --> 00:37:08.240
This means that some company sets up the
model as a model that can be hosted

00:37:08.240 --> 00:37:12.240
anywhere; they take it and host it and
give me the service to access it.

00:37:12.240 --> 00:37:16.240
I don't have to use their infrastructure,
but I have to deploy my own

00:37:16.240 --> 00:37:20.160
infrastructure, I use theirs, and they
charge me a fee for doing so.

00:37:20.160 --> 00:37:24.160
But these amounts are usually very
reasonable and allow me to use many

00:37:24.160 --> 00:37:28.040
of the applications I mentioned earlier
without having to do any deployment on my

00:37:28.040 --> 00:37:30.000
own infrastructure.

00:37:30.400 --> 00:37:33.520
Since there's very little time left,
Gregorio has already gotten up here.

00:37:33.800 --> 00:37:37.320
I'm going to mention a couple of examples
very quickly so

00:37:37.320 --> 00:37:40.320
you can see that this is much simpler
than it seems.

00:37:40.320 --> 00:37:43.880
Then, if you want, you can download the
slides and look at all the links.

00:37:43.920 --> 00:37:45.320
But I wanted to mention Civitas.

00:37:45.320 --> 00:37:48.480
Civitas is a site where people, mostly
from the computer science

00:37:48.480 --> 00:37:51.480
field, exchange models for producing
images, pictures, and video.

00:37:51.600 --> 00:37:53.680
Take a look, if you like.

00:37:53.680 --> 00:37:56.000
There is a program that everyone in that
community uses.

00:37:56.000 --> 00:37:58.760
Well, in fact there are several that
allow me to download a model and do fine

00:37:58.760 --> 00:38:00.160
tuning for certain tasks.

00:38:00.160 --> 00:38:04.160
For example, to make manga-style drawings
for me, for example, and they take models

00:38:04.160 --> 00:38:07.560
and specialize them in that and upload
the model they have generated

00:38:07.560 --> 00:38:10.560
again, so that people share models and do
things with them.

00:38:10.560 --> 00:38:14.560
And I repeat, these are people who, for
example, are very interested in

00:38:14.560 --> 00:38:18.840
image or video generation, but who do not
necessarily have computer training.

00:38:18.840 --> 00:38:22.840
And for another very interesting example,
this is how

00:38:22.840 --> 00:38:26.160
you install a model on your own
infrastructure.

00:38:26.160 --> 00:38:30.160
If you want, there's an application
called Llama or Lama, which is

00:38:30.160 --> 00:38:34.160
an interface that you install on Windows
or Linux with the command

00:38:34.160 --> 00:38:38.160
above, and with the command in the middle
you basically run it. What

00:38:38.160 --> 00:38:42.160
it gives you is a browser, a server that
you connect your browser to

00:38:42.160 --> 00:38:46.920
and you can start making inferences very
similar to how you would do with GPS.

00:38:47.200 --> 00:38:49.840
Obviously, if you don't have a GPU, if
you have a limited infrastructure, it's

00:38:49.840 --> 00:38:51.200
going

00:38:51.200 --> 00:38:54.280
to be slow, but it can be used and at
least it's good as a demo if you

00:38:54.280 --> 00:38:57.280
want to show how it works with that
specific model you have there.

00:38:57.320 --> 00:39:01.320
Gema 3D uses 31V, which is a very small 1
million parameter model

00:39:01.320 --> 00:39:04.480
compared to what we deal with now, which
is usually

00:39:04.480 --> 00:39:07.480
hundreds or thousands of millions of
parameters.

00:39:07.680 --> 00:39:10.120
However, he does things like you can talk
to him, you can say hello and he'll say,

00:39:10.120 --> 00:39:12.560
"Can I tell you a joke?" and he'll tell
it to you.

00:39:12.560 --> 00:39:15.720
Well, the joke will be like that, but he
tells it to you and you can interact with

00:39:15.720 --> 00:39:17.320
him.

00:39:17.320 --> 00:39:20.600
And really, if you look at it, you can do
that on your own computer for about ten

00:39:20.600 --> 00:39:22.280
minutes if you want.

00:39:22.280 --> 00:39:25.480
If you also have a slightly more powerful
CPU or

00:39:25.480 --> 00:39:28.480
a slightly better GPU, you can do a lot
more.

00:39:28.520 --> 00:39:31.160
The other slides contain many more
examples of this, but we don't have time

00:39:31.160 --> 00:39:32.520
to look

00:39:32.520 --> 00:39:35.840
at the treaty today, so we want you to
look at the rest of the slides.

00:39:35.840 --> 00:39:37.280
I don't know if you've noticed.

00:39:37.280 --> 00:39:40.680
So if you want to download them, and of
course, feel free to ask me any questions

00:39:40.680 --> 00:39:42.400
you may have about it.

00:39:42.400 --> 00:39:46.400
But I think the main thing, before you
kick me out, is

00:39:46.400 --> 00:39:49.600
that you believe that anyone can do this.

00:39:49.600 --> 00:39:51.440
It's much easier than it looks.

00:39:51.440 --> 00:39:55.440
The people doing this out there aren't
people with amazing technical knowledge;

00:39:55.440 --> 00:39:58.480
they're just people who like making
videos and want to find a way

00:39:58.480 --> 00:40:01.480
to make videos with artificial
intelligence, for example.

00:40:01.480 --> 00:40:04.240
It's possible. It's a matter of getting
started, and like any young person, you

00:40:04.240 --> 00:40:07.000
get into it, learn what you need to
learn, and you do it.

00:40:07.000 --> 00:40:07.920
Thank you so much.

00:40:12.400 --> 00:40:15.400
He asks himself.

00:40:15.480 --> 00:40:19.480
Okay, we have time for a comment with a
question if you want to

00:40:19.480 --> 00:40:23.480
do it, otherwise we'll leave something
for you the next day.

00:40:23.560 --> 00:40:29.480
I wanted to ask a little about Asian
models, do you know anything about them?

00:40:29.480 --> 00:40:32.440
Their infrastructure is very similar.

00:40:32.440 --> 00:40:35.280
The models that everyone is making are
very similar.

00:40:35.280 --> 00:40:39.280
I mean that each, each, each supplier,
let's say of models, usually makes

00:40:39.280 --> 00:40:43.280
some kind of change with their model in
different ways, but they don't have

00:40:43.280 --> 00:40:47.280
any special characteristics, at least not
those that come from China or

00:40:47.280 --> 00:40:50.880
Korea, which we also have some of in
Europe, most of them come, well,

00:40:50.880 --> 00:40:53.880
most, many others come from the United
States as well.

00:40:53.880 --> 00:40:56.800
Let's say that the basic architecture in
all of them is very similar.

00:40:56.800 --> 00:41:00.320
They have their own quirks, but the basic
architecture is very similar across the

00:41:00.320 --> 00:41:03.320
board, and they all basically use the
same software with minor variations.

00:41:04.400 --> 00:41:06.240
And there are more

00:41:06.240 --> 00:41:09.680
powerful in that they become.

00:41:10.200 --> 00:41:14.280
For example, using a professional point
of view.

00:41:15.160 --> 00:41:19.880
Are American models more expensive than
Asian models?

00:41:20.120 --> 00:41:23.960
They are more sorry, more than more
expensive in terms of use,

00:41:23.960 --> 00:41:26.960
which is in terms of the price they
charge you.

00:41:26.960 --> 00:41:28.880
Beautiful, you mean.

00:41:28.880 --> 00:41:32.880
Well, keep in mind that the prices, as I
was trying to explain before, the prices

00:41:32.880 --> 00:41:36.840
they are charging us are not necessarily
the cost prices, they are the prices

00:41:36.840 --> 00:41:39.840
that companies are interested in setting
so that we use them.

00:41:40.320 --> 00:41:42.960
And that implies that if, for example,
your company has a lot of capital and has

00:41:42.960 --> 00:41:44.320
committed

00:41:44.320 --> 00:41:46.960
to its shareholders to have 1,000,000,000
users next year, then it's going to

00:41:46.960 --> 00:41:48.280
provide

00:41:48.280 --> 00:41:51.280
a very cheap way to access it depending
on how you want.

00:41:51.320 --> 00:41:53.640
Even so, we found a bit of everything.

00:41:53.640 --> 00:41:56.840
There are very cheap Chinese models and
there are suppliers, in fact Chinese,

00:41:56.840 --> 00:41:58.480
that use some very cheap unlocked models.

00:41:58.480 --> 00:42:01.160
And they exist in Europe and the United
States as well.

00:42:01.160 --> 00:42:04.280
To give an example, one of the
lesser-known ones, the very

00:42:04.280 --> 00:42:07.280
cheap Nvidia, has several models that are
also open.

00:42:07.280 --> 00:42:11.280
In other words, you can download them to
their infrastructure, they cost you

00:42:11.280 --> 00:42:14.760
nothing, and you can use them up to
certain reasonable limits at zero cost.

00:42:14.760 --> 00:42:18.760
Obviously someone is paying for that, and
in this case it's Nvidia, which has its

00:42:18.760 --> 00:42:22.600
commitments to its shareholders to ensure
that it has so many thousands or millions

00:42:22.600 --> 00:42:24.560
of users next year or whatever.

00:42:25.240 --> 00:42:28.920
But remember that there is no
relationship between the cost they charge

00:42:28.920 --> 00:42:31.920
us and the cost they have to provide the
infrastructure.

00:42:33.560 --> 00:42:36.560
Thank you, thank you.

00:42:36.560 --> 00:42:39.560
Well, thank you all very much.

00:42:44.840 --> 00:42:46.760
Very good.

00:42:46.760 --> 00:42:50.760
The second presentation is by Natalia
Galván, who comes from the El Molino

00:42:50.760 --> 00:42:54.760
Special Education School in Navarra and
will talk to us about technology

00:42:54.760 --> 00:42:57.880
and pedagogy, computational thinking,
artificial

00:42:57.880 --> 00:43:00.880
intelligence at the service of inclusion,
all of that.

00:43:01.320 --> 00:43:04.320
Thank you. Okay.

00:43:04.400 --> 00:43:06.160
Let's see about this.

00:43:06.160 --> 00:43:08.440
Let's see, I think I've seen that it's
there.

00:43:08.440 --> 00:43:11.720
We have it, right? Well, thank you very
much.

00:43:12.080 --> 00:43:16.080
Good afternoon everyone, and well, first
of all, thank you for the invitation,

00:43:16.080 --> 00:43:18.720
Estefanía María, for inviting me and
bringing me here to tell you a little bit

00:43:18.720 --> 00:43:20.080
about

00:43:20.080 --> 00:43:23.080
the educational innovation project that
we discussed in a...

00:43:23.360 --> 00:43:26.440
At the school where I work, which is a
special education school called El Molino

00:43:26.440 --> 00:43:28.000
de Pamplona.

00:43:29.240 --> 00:43:33.240
So, I'll tell you about the innovation
project, how we got there, and above

00:43:33.240 --> 00:43:37.240
all, I'm interested in sharing the
activities that helped us get started

00:43:37.240 --> 00:43:40.560
in computational thinking and artificial
intelligence.

00:43:41.400 --> 00:43:45.400
But before that, I would like to share
some concerns

00:43:45.400 --> 00:43:49.040
about the current context in education.

00:43:49.040 --> 00:43:52.560
As you can see, I suppose you all agree
with me that we have increasingly

00:43:52.560 --> 00:43:55.560
diverse classrooms to which we have to
respond.

00:43:55.880 --> 00:43:59.880
And it is true that we also constantly
have very strong pressure regarding

00:43:59.880 --> 00:44:03.280
digitization, new tools, and robots that
give us all this.

00:44:03.280 --> 00:44:07.280
Ultimately, it is true that it impacts
student inclusion, because we do

00:44:07.280 --> 00:44:11.280
have to try to ensure that students are
integrated through progress,

00:44:11.280 --> 00:44:15.280
that is, that they learn, that they
participate, validate any form of

00:44:15.280 --> 00:44:17.920
participation, and that this
participation and this

00:44:17.920 --> 00:44:19.280
achievement-oriented

00:44:19.280 --> 00:44:23.280
progress makes them feel good and that
they are part of the classroom like

00:44:23.280 --> 00:44:26.440
any other student, regardless of their
diversity.

00:44:26.560 --> 00:44:30.560
So, what is the current challenge we
face, or at least the one

00:44:30.560 --> 00:44:34.560
that concerns us, at least as a center,
is how to integrate

00:44:34.560 --> 00:44:38.560
technology without creating another
barrier to student

00:44:38.560 --> 00:44:42.560
access to learning, regardless of whether
they have a disability

00:44:42.560 --> 00:44:45.960
or not? Ultimately, this is the key we
have from the

00:44:45.960 --> 00:44:48.960
perspective of our own teaching
competence.

00:44:48.960 --> 00:44:52.960
It's about creating a good instructional
design, providing support from the

00:44:52.960 --> 00:44:56.360
beginning and having a good foundation of
pedagogical intent.

00:44:56.360 --> 00:44:59.680
In other words, we want them to learn
exactly what they need to learn,

00:44:59.680 --> 00:45:02.680
not so much what tool we're going to use,
but we want them to learn.

00:45:06.480 --> 00:45:10.480
In all this you will see and agree that
in the end, as teachers, we bear a very

00:45:10.480 --> 00:45:14.480
heavy burden, because it is not only what
we often cannot control in the

00:45:14.480 --> 00:45:17.760
classroom, but also because outside we
have a digitalization that is constantly

00:45:17.760 --> 00:45:19.400
overwhelming us.

00:45:19.800 --> 00:45:23.800
So, apart from the weight we have and the
role we have as mediators between

00:45:23.800 --> 00:45:27.320
technology and the students we have in
the classroom, we are indeed

00:45:27.320 --> 00:45:30.320
impacted by being able to make a learning
situation accessible.

00:45:30.360 --> 00:45:34.360
In other words, all students in the
classroom, regardless of

00:45:34.360 --> 00:45:38.360
their personal circumstances, can access,
express themselves

00:45:38.360 --> 00:45:42.360
and participate, not under the universal
design of learning.

00:45:43.240 --> 00:45:47.240
And all of this through a pedagogical
intention, as I said before, of the

00:45:47.240 --> 00:45:50.280
digital tools that we have or think are
the most appropriate

00:45:50.280 --> 00:45:53.280
for the point at which the students I
have in question are.

00:45:53.720 --> 00:45:57.720
Obviously, we're not all going to take on
the burden, but we also have many

00:45:57.720 --> 00:46:01.160
contextual barriers that will depend
somewhat on the type of

00:46:01.160 --> 00:46:04.160
center, what we're doing, and the
training we have.

00:46:04.200 --> 00:46:07.280
So, given this situation, how do we do
it?

00:46:09.160 --> 00:46:13.160
Well, for us, things did get a bit of a
breakthrough when,

00:46:13.160 --> 00:46:17.160
through computational thinking and under
that STEM approach,

00:46:17.160 --> 00:46:21.160
we saw that it was the starting point for
any digital learning

00:46:21.160 --> 00:46:25.160
with students and that, in turn, it would
allow us to work on

00:46:25.160 --> 00:46:29.160
the development of digital competence,
because today it

00:46:29.160 --> 00:46:34.600
has become a way of participating more,
almost in many cases the only way.

00:46:35.560 --> 00:46:39.560
And in the students we had, we did see
that in the end,

00:46:39.560 --> 00:46:43.560
working on computational thinking allowed
us to work

00:46:43.560 --> 00:46:47.560
with them on the reflective process of
learning to learn,

00:46:47.560 --> 00:46:51.560
which is fundamental not only for
learning artificial

00:46:51.560 --> 00:46:56.960
intelligence, but also for making it
transferable to daily life, to autonomy.

00:46:56.960 --> 00:47:00.960
In all this sense, we saw that
computational thinking was becoming a

00:47:00.960 --> 00:47:04.240
powerful tool for students, at whatever
level they are, to exercise active

00:47:04.240 --> 00:47:05.920
citizenship.

00:47:05.920 --> 00:47:11.560
And as mentioned at the beginning,
through logical and critical thinking.

00:47:14.320 --> 00:47:15.920
We're talking about computational
thinking.

00:47:15.920 --> 00:47:18.160
But how does artificial intelligence
work?

00:47:18.160 --> 00:47:20.560
How do we fit him into all of this, or
what does he have to do with it?

00:47:20.560 --> 00:47:24.720
Well, the European Digital Competence
Framework has recently been published.

00:47:25.240 --> 00:47:28.600
This one you have here is from the 5th
edition of last year.

00:47:28.680 --> 00:47:32.320
And well, you see that it talks about
five skills that you have listed there.

00:47:32.360 --> 00:47:34.560
Information search, evaluation and
management, communication and

00:47:34.560 --> 00:47:35.680
collaboration.

00:47:35.720 --> 00:47:40.600
Content creation, safety, well-being and
responsible use, and problem-solving.

00:47:40.600 --> 00:47:43.120
These five dimensions or competencies
have been transformed into 21

00:47:43.120 --> 00:47:44.400
competencies

00:47:44.400 --> 00:47:47.400
in which I don't know if you see it
there, but the little arrow.

00:47:47.400 --> 00:47:53.240
This computational thinking has become a
fundamental skill with its own identity.

00:47:53.400 --> 00:47:57.400
And the novelty is that in the end,
artificial intelligence is becoming a

00:47:57.400 --> 00:48:01.400
completely cross-cutting skill among the
rest, and it basically focuses

00:48:01.400 --> 00:48:05.400
on the fact that we have to educate our
students in data processing or in

00:48:05.400 --> 00:48:08.760
knowing what data is, and above all, in
interacting with AI in a critical and

00:48:08.760 --> 00:48:10.480
responsible way.

00:48:10.800 --> 00:48:14.320
Well, I don't know if you can see it, but
the QR code was left there.

00:48:16.160 --> 00:48:19.000
Okay, so, how else not?

00:48:19.000 --> 00:48:23.000
For example, in my case with students
with intellectual disabilities,

00:48:23.000 --> 00:48:26.920
how do we integrate computational
thinking, artificial intelligence, and

00:48:26.920 --> 00:48:28.880
these special needs?

00:48:29.600 --> 00:48:33.600
Well, as I was saying before, through
computational thinking, through the

00:48:33.600 --> 00:48:37.600
development of computational thinking
skills, we are talking about

00:48:37.600 --> 00:48:41.600
decomposition, pattern recognition,
abstraction, algorithmic thinking,

00:48:41.600 --> 00:48:45.440
and error evaluation and debugging, which
allow us to work with many of them on the

00:48:45.440 --> 00:48:47.400
development of executive functions.

00:48:47.400 --> 00:48:51.400
It is true that there is evidence and
studies that confirm that many

00:48:51.400 --> 00:48:55.400
of the profiles we have in attention to
diversity do show many

00:48:55.400 --> 00:48:59.400
challenges in these functions. They will
allow us to work on

00:48:59.400 --> 00:49:03.400
metacognition, not always at the level
they have, but so that they

00:49:03.400 --> 00:49:06.920
are able to reflect on their own learning
and therefore, the development of

00:49:06.920 --> 00:49:08.680
adaptive skills.

00:49:10.280 --> 00:49:14.280
And on the other hand, we approach the
concept of artificial intelligence,

00:49:14.280 --> 00:49:18.280
understanding the algorithm as a learning
sequence, as a sequence of

00:49:18.280 --> 00:49:21.880
itself, as a routine that will ultimately
allow it to access

00:49:21.880 --> 00:49:24.880
learning and be more autonomous and able
to participate.

00:49:25.160 --> 00:49:29.160
In our case, we also approach it from one
of the branches of artificial

00:49:29.160 --> 00:49:33.160
intelligence, which is machine learning,
for which a manipulative or very

00:49:33.160 --> 00:49:36.880
costly task is fundamental and which is
not usually liked, or at least we were

00:49:36.880 --> 00:49:38.760
told that it is the data one.

00:49:39.480 --> 00:49:43.360
We're talking about classifying and
labeling a dataset and reviewing

00:49:43.360 --> 00:49:46.360
it according to the instructions they
give us. They are.

00:49:46.400 --> 00:49:50.080
These are structured, repetitive tasks
with very strict rules and very little

00:49:50.080 --> 00:49:51.920
cognitive load.

00:49:52.240 --> 00:49:56.240
Ultimately, it's a very basic level of
Bloom's

00:49:56.240 --> 00:50:00.240
taxonomy, and as I said, these are all
the machine

00:50:00.240 --> 00:50:03.280
learning models we have available today.

00:50:03.280 --> 00:50:07.240
These tasks are required in order to
train the model, which is why it

00:50:07.240 --> 00:50:10.240
is within the artificial intelligence
value chain.

00:50:10.240 --> 00:50:14.240
It is a very powerful value and that is
where we saw a niche to be able to start

00:50:14.240 --> 00:50:18.640
with the social and labor integration of
students with intellectual disabilities.

00:50:20.600 --> 00:50:24.600
Consequently, we have been able to see
how we can generate a

00:50:24.600 --> 00:50:28.600
new professional competence, which we
have named in our

00:50:28.600 --> 00:50:32.280
context, which is the professional
competence in auxiliary tasks and data

00:50:32.280 --> 00:50:34.120
review.

00:50:34.120 --> 00:50:37.320
And this has been the final result of the
educational

00:50:37.320 --> 00:50:40.320
innovation project that we have carried
out in my school.

00:50:40.640 --> 00:50:42.760
Well, as I was saying, I come from
Pamplona.

00:50:42.760 --> 00:50:45.400
That's a picture of my school, which is
very close to the old town. It's a

00:50:45.400 --> 00:50:46.760
special

00:50:46.760 --> 00:50:50.520
education school, subsidized by the
Government of Navarre, and belongs to the

00:50:50.520 --> 00:50:52.400
Ferrer Citizen Foundation.

00:50:52.400 --> 00:50:56.400
We have an educational service, an
occupational service,

00:50:56.400 --> 00:51:00.400
a Family Guidance Center and a Sports
Club, which is also

00:51:00.400 --> 00:51:03.640
very important and currently has 76
students.

00:51:03.640 --> 00:51:06.280
You can see that we have agreements with
the stages from Early Childhood Education

00:51:06.280 --> 00:51:07.640
to Special

00:51:07.640 --> 00:51:10.280
Vocational Training, and as I was saying,
we have turned that professional

00:51:10.280 --> 00:51:11.640
competence into a

00:51:11.640 --> 00:51:14.280
specialization course that the Department
of Education has told us about, which is

00:51:14.280 --> 00:51:15.640
the Specialization

00:51:15.640 --> 00:51:18.280
Course in Artificial Intelligence for
people with intellectual disabilities who

00:51:18.280 --> 00:51:19.640
have just

00:51:19.640 --> 00:51:22.280
entered the system, and I don't know
what's happening at the parent company,

00:51:22.280 --> 00:51:23.640
but at least in Navarre

00:51:23.640 --> 00:51:26.720
they finish the app and there is very
little accessible training for people

00:51:26.720 --> 00:51:28.280
with intellectual disabilities.

00:51:28.400 --> 00:51:32.120
Well, you see there it's for people over
22 and it lasts for a year.

00:51:34.320 --> 00:51:34.920
Okay.

00:51:34.920 --> 00:51:36.840
How did I get into this specialization
course?

00:51:36.840 --> 00:51:40.840
Well, through the educational innovation
project that I mentioned, which we

00:51:40.840 --> 00:51:44.280
started in the 21-22 school year, that
is, just after the pandemic.

00:51:44.280 --> 00:51:48.280
We got quite a scare when they told us
that they were proposing that our school

00:51:48.280 --> 00:51:52.280
participate in the Artificial
Intelligence week for educational centers

00:51:52.280 --> 00:51:56.280
that the Navarre government is piloting,
and also, precisely, we were starting

00:51:56.280 --> 00:52:00.280
with a new education law and through the
Lomloe they were already explicitly

00:52:00.280 --> 00:52:03.840
talking about computational thinking in
the socio-labor field.

00:52:04.120 --> 00:52:06.760
Our foundation, to which we belong as a
school, was beginning to organize itself

00:52:06.760 --> 00:52:08.120
together with

00:52:08.120 --> 00:52:10.880
another Navarrese entity, which is in the
same vein, that has special employment

00:52:10.880 --> 00:52:13.640
centers for people with intellectual
disabilities.

00:52:13.760 --> 00:52:16.400
They were beginning to form the Teresa
Foundation, which is a foundation that

00:52:16.400 --> 00:52:17.760
tries to find

00:52:17.760 --> 00:52:21.040
employment niches in technology for
people with intellectual disabilities.

00:52:21.040 --> 00:52:24.040
Therefore, the need was there.

00:52:24.040 --> 00:52:28.040
We had to respond to a training demand
for employment for people

00:52:28.040 --> 00:52:33.240
with intellectual disabilities in the
digital field and digital technologies.

00:52:35.640 --> 00:52:36.880
So how do we begin?

00:52:36.880 --> 00:52:38.240
Now it's true, five years ago.

00:52:38.240 --> 00:52:41.520
It may not seem like that much time has
passed, but I remember five years ago

00:52:41.520 --> 00:52:43.160
when they came to tell us about this.

00:52:43.920 --> 00:52:46.560
I know that some of my colleagues in the
force would have run away and said,

00:52:46.560 --> 00:52:47.920
'We're done with this,

00:52:47.920 --> 00:52:51.280
we don't want to know anything about it,'
because it wasn't that loud anyway.

00:52:51.320 --> 00:52:53.960
We were just starting to explore the
topic of artificial intelligence; we

00:52:53.960 --> 00:52:55.320
didn't know where to

00:52:55.320 --> 00:52:58.240
begin. We saw that there was
computational thinking, the curriculum,

00:52:58.240 --> 00:52:59.720
but no, we had no idea where to start.

00:52:59.720 --> 00:53:03.720
So, well, two of us colleagues were
training in computational thinking

00:53:03.720 --> 00:53:07.720
in order to later replicate the training
to the faculty and to start

00:53:07.720 --> 00:53:11.720
generating preliminary activities that
would prepare the students for

00:53:11.720 --> 00:53:15.720
the AI ​​week, that we were going to have
an AI week, that we had clear

00:53:15.720 --> 00:53:19.400
objectives, that the students would have
an active role that they would leave.

00:53:19.440 --> 00:53:22.560
It's great to watch YouTube videos and
have them, but...

00:53:22.560 --> 00:53:26.560
But we wanted them to be technology
generators with very

00:53:26.560 --> 00:53:31.560
basic introductory concepts of what an
algorithm is and block programming.

00:53:31.560 --> 00:53:33.520
And also very importantly, with our
students.

00:53:33.520 --> 00:53:37.520
In terms of cybersecurity, the students
we have at the

00:53:37.520 --> 00:53:41.320
school are very vulnerable to any scam,
anything at all.

00:53:41.360 --> 00:53:44.000
So, well, we felt the issue of fake
accounts and digital identity was

00:53:44.000 --> 00:53:45.320
fundamental.

00:53:45.320 --> 00:53:48.360
And we developed all of this through
digital applications, primarily with

00:53:48.360 --> 00:53:49.920
Code.

00:53:50.080 --> 00:53:53.000
Well, if you don't know, well, I'm sure
you do, okay.

00:53:53.000 --> 00:53:56.840
This is a screenshot of the Symbaloo we
prepared with these activities for each

00:53:56.840 --> 00:53:58.800
pair of students.

00:53:59.640 --> 00:54:01.360
This one, this one is bad.

00:54:01.360 --> 00:54:05.240
It was dedicated to the students of
Special Vocational Training

00:54:05.240 --> 00:54:08.240
and it was, as you can see, very
sequenced at the end.

00:54:08.240 --> 00:54:11.240
We started with a simple code course.

00:54:11.320 --> 00:54:14.360
There was something for every skill, even
just getting close to the screen and

00:54:14.360 --> 00:54:15.880
dragging from there.

00:54:15.880 --> 00:54:19.880
We were working on digital skills to then
move on,

00:54:19.880 --> 00:54:23.640
and that was scaling up to small block
programs.

00:54:24.040 --> 00:54:24.600
Then it scales.

00:54:24.600 --> 00:54:27.440
Let's go to the concept of AI through a
very simple course they have about and

00:54:27.440 --> 00:54:30.280
for the oceans, which is classifying
garbage.

00:54:30.840 --> 00:54:34.840
And then we were already getting into the
cybersecurity part,

00:54:34.840 --> 00:54:37.920
mainly trying to decrypt, but not so much
anymore.

00:54:37.960 --> 00:54:40.480
We wanted them to decipher it, but oh
well, it's an app.

00:54:40.480 --> 00:54:43.120
The phase is an application that shows
you two images, and we were trying to

00:54:43.120 --> 00:54:44.480
play a game to

00:54:44.480 --> 00:54:47.280
see if he could tell which one was made
by AI and which one wasn't, simply to

00:54:47.280 --> 00:54:50.120
raise Andrew's awareness, since he
skipped it.

00:54:50.160 --> 00:54:54.160
Talking a little about pattern
recognition and machine

00:54:54.160 --> 00:54:57.880
learning to understand how they could
quickly create their own machine learning

00:54:57.880 --> 00:54:59.760
model.

00:55:01.200 --> 00:55:05.200
Well, here you see some images and it is
true that AI week was initially geared

00:55:05.200 --> 00:55:08.440
towards older students, those in
vocational training.

00:55:08.640 --> 00:55:12.640
Yes, we are very proud because we managed
to get it through all

00:55:12.640 --> 00:55:16.320
the educational stages, from the
preschool you see there, to secondary

00:55:16.320 --> 00:55:18.200
school, etc.

00:55:19.840 --> 00:55:23.400
It was also a very enriching week
because, in addition to student training,

00:55:23.400 --> 00:55:26.400
training was also provided to families
and teachers.

00:55:26.960 --> 00:55:30.960
And then it also came with a very
interesting exhibition, a bit

00:55:30.960 --> 00:55:34.960
about the key concepts of artificial
intelligence and about the

00:55:34.960 --> 00:55:38.400
people who had emerged with this
technology, which was covered in the

00:55:38.400 --> 00:55:40.160
local press.

00:55:40.160 --> 00:55:43.440
The truth is that it had a big impact and
it was a week of a lot of

00:55:43.440 --> 00:55:46.440
learning, a lot of work, but one that was
very enjoyable.

00:55:48.240 --> 00:55:51.680
AI Week has passed and now we're
wondering what to do next?

00:55:51.680 --> 00:55:55.680
Well, we launched an evaluation for all
the teachers so they could give us some

00:55:55.680 --> 00:55:58.680
feedback, things to improve or what we
could do from now on.

00:55:59.280 --> 00:56:03.280
And well, we carried out another training
that we felt was important on

00:56:03.280 --> 00:56:06.400
curriculum integration and mathematical
development.

00:56:06.440 --> 00:56:09.760
So, computational thinking is fine, but
where's the curriculum?

00:56:10.280 --> 00:56:13.960
And then it is true that we felt it was
important to be able to continue a little

00:56:13.960 --> 00:56:15.800
with the work we had done during AI week.

00:56:15.800 --> 00:56:19.800
Systematize at all stages, at all, from
early childhood to

00:56:19.800 --> 00:56:25.040
small sessions of computational thinking
and artificial intelligence.

00:56:25.920 --> 00:56:29.920
Therefore, it was important for us to
understand where we saw computational

00:56:29.920 --> 00:56:33.920
thinking. For our part, well, from the
mathematical field and the field of

00:56:33.920 --> 00:56:36.560
knowledge of the doctor of the
environment, sorry, the social field, in

00:56:36.560 --> 00:56:37.920
order

00:56:37.920 --> 00:56:41.200
to work on digital competence and
competence, learning to learn to learn

00:56:41.200 --> 00:56:42.880
personal and social.

00:56:42.880 --> 00:56:46.280
Regarding everything I was telling you,
we put it on a weekly or

00:56:46.280 --> 00:56:49.280
bi-weekly schedule depending on the needs
of the group.

00:56:49.320 --> 00:56:53.320
And we did think it was important that,
as I mentioned before, for there

00:56:53.320 --> 00:56:57.320
to be a real transfer, a positive and
tangible transfer, they would do

00:56:57.320 --> 00:57:01.320
something, that is, not just be on the
screen, but in the end they had to

00:57:01.320 --> 00:57:05.320
create some kind of mural or some kind of
classification game and through

00:57:05.320 --> 00:57:09.320
the transversality of key content
elements that we work on at all stages,

00:57:09.320 --> 00:57:13.320
such as money management, shopping, road
safety education and centers

00:57:13.320 --> 00:57:16.360
of interest specific to the students.

00:57:18.040 --> 00:57:22.040
And yes, we do focus a lot on
mathematical thinking, because in the end

00:57:22.040 --> 00:57:26.040
our students are in different phases of
learning and we saw that the

00:57:26.040 --> 00:57:29.960
development of this thinking is the
earliest, through games and the senses of

00:57:29.960 --> 00:57:31.920
sensory development, you know.

00:57:31.960 --> 00:57:35.520
And that the environment ultimately acts
as a source of learning for all of them.

00:57:35.920 --> 00:57:37.040
Continuously.

00:57:37.040 --> 00:57:41.040
From a young age, we are constantly
creating mental

00:57:41.040 --> 00:57:45.040
representations that we ultimately use to
try to observe

00:57:45.040 --> 00:57:49.040
and understand the world around us, to
solve everyday

00:57:49.040 --> 00:57:53.040
situations, and this learning process we
undertake

00:57:53.040 --> 00:57:57.800
with these mental representations
generates autonomous thinking.

00:57:57.800 --> 00:58:01.800
For all of this, and as I mentioned,
through the development of skills,

00:58:01.800 --> 00:58:05.320
especially executive functions, and
that's a little bit where we start.

00:58:06.800 --> 00:58:10.800
We talked about logical reasoning, number
measurement,

00:58:10.800 --> 00:58:16.640
problem solving, geometric knowledge, and
representation and communication.

00:58:18.240 --> 00:58:21.360
And yes, we did separate.

00:58:21.360 --> 00:58:25.360
As I mentioned, in the end we had
students at different learning levels

00:58:25.360 --> 00:58:28.720
and we did make a sort of classification
of the students we had.

00:58:28.720 --> 00:58:31.360
On one hand, we had students who were
still in a very early stage of learning,

00:58:31.360 --> 00:58:32.720
and for them

00:58:32.720 --> 00:58:36.440
we were going to dedicate some activities
called prerequisites, which in the end

00:58:36.440 --> 00:58:39.440
is working on those cognitive and
mathematical thinking skills.

00:58:39.480 --> 00:58:43.480
And then another one that you will see
later is computational thinking

00:58:43.480 --> 00:58:47.480
as such, with artificial intelligence for
the three requirements

00:58:47.480 --> 00:58:51.480
of typical activities that we start from
the applications of cause

00:58:51.480 --> 00:58:55.480
and effect, series, recognition,
identification, classification,

00:58:55.480 --> 00:58:59.480
categorization, sequencing, toys,
constructions, tangram, word

00:58:59.480 --> 00:59:03.480
searches, mazes, Connect the Dots, Find
the Differences, Inferences,

00:59:03.480 --> 00:59:07.720
Intruder, Memory, lotto, Body Games, Not
through imitation.

00:59:07.720 --> 00:59:10.720
Table Tiles.

00:59:11.040 --> 00:59:11.920
Little technology.

00:59:11.920 --> 00:59:14.480
You can see what is needed to start
working on that mathematical thinking,

00:59:14.480 --> 00:59:15.760
that

00:59:15.760 --> 00:59:18.760
computational thinking that is the basis
of any digital technology.

00:59:20.240 --> 00:59:24.240
A brief overview: In the end, these
activities were intended for children

00:59:24.240 --> 00:59:27.640
and, in fact, one of the primary school
students and the transition to adulthood

00:59:27.640 --> 00:59:29.360
program.

00:59:29.360 --> 00:59:33.320
Older people with greater needs, with
whom we started a little bit who might be

00:59:33.320 --> 00:59:35.320
able to help us.

00:59:35.760 --> 00:59:39.760
We downloaded this application, which is
free on the CREA website,

00:59:39.760 --> 00:59:43.760
which is the Resource Center for
Educational Equity of Navarre,

00:59:43.760 --> 00:59:46.840
meaning it is freely available for
download.

00:59:46.840 --> 00:59:50.160
And it's a great app because it also
allows you

00:59:50.160 --> 00:59:53.160
to work on any type of educational area.

00:59:53.320 --> 00:59:56.680
And there is a part that does have
computational thinking.

00:59:57.680 --> 01:00:01.680
We also use the Pica app a lot, which is
for iOS and

01:00:01.680 --> 01:00:05.680
because it allows you to customize peer
association,

01:00:05.680 --> 01:00:09.320
categorization, and sequencing activities

01:00:09.320 --> 01:00:12.320
with photos of the student or with
objects.

01:00:12.320 --> 01:00:15.520
They allow you to personalize them with
very specific things that we know will

01:00:15.520 --> 01:00:17.120
ultimately motivate the students.

01:00:17.120 --> 01:00:18.200
And then with Symbaloo.

01:00:18.200 --> 01:00:21.480
There you can see a bit of the structure
from one to nine or from one to whatever

01:00:21.480 --> 01:00:23.160
it may be.

01:00:23.160 --> 01:00:28.800
Each item in the symbol directly opens to
a digital resource.

01:00:28.840 --> 01:00:32.840
So, as I was saying, with activities
similar to those places

01:00:32.840 --> 01:00:35.880
we've been talking about, we ultimately
try to

01:00:35.880 --> 01:00:38.880
work on that autonomous learning of the
students.

01:00:40.840 --> 01:00:44.840
For older students, between the "I owe
one and I have two" stage, which

01:00:44.840 --> 01:00:48.840
is primary and secondary school, and with
the Vocational Training

01:00:48.840 --> 01:00:52.840
cycle, for computational thinking we did
about four categories of

01:00:52.840 --> 01:00:57.440
activities: plugged-in floor board and
table board activities, also on paper.

01:00:58.120 --> 01:01:02.120
And then there were some that were
directly with robots

01:01:02.120 --> 01:01:06.120
and the panel that came with the robot, a
tabletop directly

01:01:06.120 --> 01:01:10.120
to the computer via tablet or software
with applications,

01:01:10.120 --> 01:01:14.920
and then joining the two parts, the robot
and the computer.

01:01:17.360 --> 01:01:21.120
Okay, here you can see a small example in
all classrooms, and that's important.

01:01:21.240 --> 01:01:23.880
Then we got scolded quite a bit, but
well, in all the classrooms we painted

01:01:23.880 --> 01:01:25.240
the

01:01:25.240 --> 01:01:28.480
grids with a stamp, which I imagine
you're already saying, but

01:01:28.480 --> 01:01:31.480
in all of them we painted that grid to
the scale we could.

01:01:32.120 --> 01:01:36.120
So, we categorized them as movement
because in the end it was the

01:01:36.120 --> 01:01:40.120
part that most motivated the students,
because we started

01:01:40.120 --> 01:01:44.120
moving chairs and tables and we started
moving around and working

01:01:44.120 --> 01:01:48.040
on the grid that became the game board,
the code we had to be able

01:01:48.040 --> 01:01:51.040
to move on that game board that you can
see.

01:01:51.040 --> 01:01:55.040
The arrows are directions from top to
bottom, right,

01:01:55.040 --> 01:01:59.040
left, so that they could then build the
sequence of steps

01:01:59.040 --> 01:02:02.840
or the algorithm to achieve any challenge
we proposed.

01:02:02.840 --> 01:02:05.840
On the game board in this area, the.

01:02:07.640 --> 01:02:11.640
The content we worked on depended a lot
on the program each group

01:02:11.640 --> 01:02:15.640
was in, but well, we worked on the solar
system, money management,

01:02:15.640 --> 01:02:19.640
recycling, vocabulary, skills and healthy
habits and in the

01:02:19.640 --> 01:02:23.640
end the role that one had was the
programmer and another was the

01:02:23.640 --> 01:02:27.640
robot and the others who were in class
had to write down the code

01:02:27.640 --> 01:02:31.640
that was written or debug the errors of
the classmate who was

01:02:31.640 --> 01:02:35.640
programming or another who was
classifying a little or

01:02:35.640 --> 01:02:40.560
completing that final product that they
had to do and various support.

01:02:40.640 --> 01:02:44.000
Each student, each student, needed a
different kind of support,

01:02:44.000 --> 01:02:47.000
which could be visual through oral
gestural code.

01:02:47.040 --> 01:02:50.040
Well, we were going, they were learning,
but so were we.

01:02:51.320 --> 01:02:55.320
This is another example of a moving
activity where in

01:02:55.320 --> 01:02:59.320
this case we had a student who, instead
of simply leaving

01:02:59.320 --> 01:03:03.320
the code on the floor, needed to put the
code directly

01:03:03.320 --> 01:03:08.520
onto the grid, or else the tutor would
support him in building the map.

01:03:13.200 --> 01:03:16.000
It's the same with the older ones, isn't
it?

01:03:16.000 --> 01:03:18.360
A bit of the same scheme and the
opposite.

01:03:18.360 --> 01:03:22.360
The truth is that in the end it allowed
you to work on a lot of content,

01:03:22.360 --> 01:03:27.120
and also here in the end what is needed
is a lot of creativity and a lot of time.

01:03:27.120 --> 01:03:31.120
Sometimes, in order to adapt everything
and see a little bit where each

01:03:31.120 --> 01:03:34.160
group is at, we can give them meaningful
activities.

01:03:34.200 --> 01:03:38.200
But well, under the scheme of my
programming code, the type of dynamics

01:03:38.200 --> 01:03:42.200
we did and a kind of challenge-solving,
the truth is that they enjoyed

01:03:42.200 --> 01:03:46.200
it a lot, they were getting hooked,
getting more and more involved

01:03:46.200 --> 01:03:50.200
because we chose that code that you see
there in the background of the

01:03:50.200 --> 01:03:54.200
arrows, because in the end we saw that it
was the most significant

01:03:54.200 --> 01:03:58.200
transfer that they will be able to make,
for example, a visual

01:03:58.200 --> 01:04:02.200
programming like Scratch and before, if
not to the visual programming

01:04:02.200 --> 01:04:05.720
that we use in Code, which I suppose you
know are arrows that indicate arrows and

01:04:05.720 --> 01:04:07.520
so on.

01:04:07.520 --> 01:04:11.520
So, the idea was to start from scratch
and then be able to

01:04:11.520 --> 01:04:16.040
scale up to a more technological
programming environment.

01:04:16.640 --> 01:04:20.640
We also made several of these cards which
are essentially

01:04:20.640 --> 01:04:24.200
mazes that you have to overcome obstacles
to get, in

01:04:24.200 --> 01:04:27.200
this case cars, and allow you to write
the code.

01:04:27.200 --> 01:04:30.880
It is true that it allows for a more
reflective approach, but I'm telling you,

01:04:30.880 --> 01:04:32.720
it doesn't really impress us.

01:04:32.760 --> 01:04:35.760
In other words, we gave out tokens and

01:04:35.960 --> 01:04:38.960
little interest.

01:04:39.120 --> 01:04:41.160
Okay, back

01:04:41.160 --> 01:04:44.120
to review.

01:04:44.120 --> 01:04:46.080
Forward

01:04:46.080 --> 01:04:48.000
It is land

01:04:48.000 --> 01:04:49.280
forward.

01:04:49.280 --> 01:04:51.560
Don't you want to?

01:04:51.560 --> 01:04:55.560
Well, in the end we involved the whole
center and even the Physical Education

01:04:55.560 --> 01:04:59.560
staff also work on computational thinking
in the simplest way, instead of using

01:04:59.560 --> 01:05:03.280
a grid with hoops to work on the whole
issue of body schema and laterality.

01:05:05.200 --> 01:05:08.600
In primary school, I am already, for
example, in the example of a classroom

01:05:08.600 --> 01:05:10.320
with an ASD profile.

01:05:10.880 --> 01:05:14.880
The game board was already on the table
with another robot,

01:05:14.880 --> 01:05:19.320
but one of those simple ones that are
like floor-standing.

01:05:19.360 --> 01:05:23.360
And it is true that here the adaptation
or adjustment we made was to give

01:05:23.360 --> 01:05:27.360
the code through social history, which in
the end also, because it is not

01:05:27.360 --> 01:05:31.040
so much that many times they cannot do
it, but that they can do it and we have

01:05:31.040 --> 01:05:32.920
to mediate a little there.

01:05:32.960 --> 01:05:36.960
Then we saw that in the end it simply
presents a social

01:05:36.960 --> 01:05:40.840
story with the color code and the typical
123 sequence.

01:05:41.000 --> 01:05:45.000
The students quickly understood how the
robot worked through direct modeling,

01:05:45.000 --> 01:05:49.080
and in this way they could understand
what activity was being proposed to them.

01:05:49.280 --> 01:05:53.400
In addition, it will also allow us to
work on social skills.

01:05:53.400 --> 01:05:56.400
Finally, the turn, the wait.

01:05:56.600 --> 01:06:00.280
And the truth is that this was a very,
very interesting project

01:06:00.280 --> 01:06:03.280
and it was very, very well received by
the students.

01:06:04.000 --> 01:06:06.400
And now, moving on to the older ones,
well, this is what I was telling you

01:06:06.400 --> 01:06:07.600
about

01:06:07.600 --> 01:06:10.600
in the first photo, you can see the Codec
programming there.

01:06:10.600 --> 01:06:14.480
So, is it ultimately a visual program
with arrows to create the

01:06:14.480 --> 01:06:17.480
scale I mentioned earlier in the middle?
See?

01:06:17.480 --> 01:06:21.480
Maite is doing an activity that
compresses, which I also suggested for

01:06:21.480 --> 01:06:25.280
lower levels, but which can also be
worked on in the end according to the

01:06:25.280 --> 01:06:27.200
pedagogical sense we give it.

01:06:27.200 --> 01:06:30.800
It is possible to work with other needs.

01:06:31.720 --> 01:06:35.720
We also introduced the Blue bot robot,
which is very similar to the

01:06:35.720 --> 01:06:39.720
vivo one that I imagine you know, but
with a slightly more adult aspect

01:06:39.720 --> 01:06:43.640
that is sometimes difficult to find,
allowing us to work from a tablet,

01:06:43.640 --> 01:06:45.640
computer, and tabletop.

01:06:46.280 --> 01:06:50.280
Here's another student, starting to work
with Minecraft, with

01:06:50.280 --> 01:06:55.160
some much more complicated programming,
and another student with picto blogs.

01:06:55.200 --> 01:06:58.360
Well, in the end we were pretty much
doing what I was saying, they were

01:06:58.360 --> 01:06:59.960
showing us the way.

01:07:00.760 --> 01:07:03.400
It wasn't until this first year that we
were working on this educational

01:07:03.400 --> 01:07:04.760
innovation

01:07:04.760 --> 01:07:07.400
project that we managed to formulate our
own curriculum for the Vocational

01:07:07.400 --> 01:07:08.760
Training

01:07:08.760 --> 01:07:12.360
cycle specifically for the Artificial
Intelligence module, which is what we

01:07:12.360 --> 01:07:14.160
implemented the following year.

01:07:16.760 --> 01:07:17.840
On the 22nd and 23rd.

01:07:17.840 --> 01:07:20.320
We've already started this new module in
the vocational training program, which

01:07:20.320 --> 01:07:21.600
has been very successful.

01:07:21.600 --> 01:07:25.600
We decided to participate in the Science
and Technology Fair, which

01:07:25.600 --> 01:07:30.720
is a fair held in Navarre, with different
schools to showcase their STEM projects.

01:07:30.720 --> 01:07:34.720
And we also felt it was important to give
visibility to the work of people with

01:07:34.720 --> 01:07:38.840
intellectual disabilities, who should not
have a place in this framework.

01:07:39.040 --> 01:07:43.040
And we managed to extend this integration
that we had made in the app to

01:07:43.040 --> 01:07:46.600
the rest of the educational stages in the
23-24 school year.

01:07:46.600 --> 01:07:50.600
Seeing the great success, the great
response from the students, we launched

01:07:50.600 --> 01:07:53.600
the educational suggestion to create the
specialization

01:07:53.600 --> 01:07:56.600
course that I mentioned earlier, a
specialization course.

01:07:56.600 --> 01:08:00.600
Because it is true, the normal vocational
training, misnamed, was

01:08:00.600 --> 01:08:04.600
falling, it does have the option for
students to take specialized

01:08:04.600 --> 01:08:08.600
courses, so why won't people with
intellectual disabilities also

01:08:08.600 --> 01:08:12.040
have the option of having a specialized
training course?

01:08:12.200 --> 01:08:15.720
Well, we launched the course and it was
accepted by the Department of Education.

01:08:15.720 --> 01:08:18.360
We are very, very grateful and it was
generated through the Department of

01:08:18.360 --> 01:08:19.680
Education.

01:08:19.880 --> 01:08:23.880
A network of computational thinking

01:08:23.880 --> 01:08:28.280
colleges was created, and we are part of
it.

01:08:28.840 --> 01:08:32.520
Last year we continued with the second
edition of the

01:08:32.520 --> 01:08:35.520
Specialization Course in Artificial
Intelligence.

01:08:35.520 --> 01:08:39.520
We have introduced 3D printing, which is
also a marvel how they work

01:08:39.520 --> 01:08:43.520
with 3D printing at the end of
computational thinking, sequencing

01:08:43.520 --> 01:08:47.520
the instructions within the network that
I mentioned of educational

01:08:47.520 --> 01:08:50.800
centers, a working network has been
generated for special

01:08:50.800 --> 01:08:53.800
educational needs and yes we launch many.

01:08:53.800 --> 01:08:57.800
We have been providing training to other
teachers in the

01:08:57.800 --> 01:09:01.800
community, we have had the opportunity to
participate in

01:09:01.800 --> 01:09:05.800
dissemination days and well, this year
that we are already

01:09:05.800 --> 01:09:09.800
immersed in the third edition of the
course, in a month we

01:09:09.800 --> 01:09:12.880
will have the AI ​​week that has been
created.

01:09:13.200 --> 01:09:16.720
We have received the approval of the
Department of Education

01:09:16.720 --> 01:09:19.720
to give us the option of designing it to
our liking.

01:09:20.200 --> 01:09:23.280
And well, and as a new development, this
year we finished in partnership with the

01:09:23.280 --> 01:09:24.840
Government of Navarre.

01:09:24.840 --> 01:09:27.480
We have started a new project called
Inclusiva, which, like our specialization

01:09:27.480 --> 01:09:28.840
course,

01:09:28.840 --> 01:09:32.040
will allow our students to do internships
in special employment and

01:09:32.040 --> 01:09:35.040
occupational centers that have this
technological focus.

01:09:35.840 --> 01:09:39.840
Our commitment through this project,
which is a cross-border cooperation

01:09:39.840 --> 01:09:43.640
project with France, is to enable our
students to take the leap to

01:09:43.640 --> 01:09:46.640
real company internships, with support
and more.

01:09:46.640 --> 01:09:49.640
But hey, that's the big challenge we
have.

01:09:50.240 --> 01:09:54.240
But aside from all these milestones, from
this whole path we've

01:09:54.240 --> 01:09:58.080
been developing in educational
innovation, with computational

01:09:58.080 --> 01:10:01.080
thinking and AI, does it really have all
this?

01:10:01.080 --> 01:10:05.080
It is successful, it has value when you
see that there are two people like Ana

01:10:05.080 --> 01:10:09.080
and there are not two students who have
had the specialization course, who

01:10:09.080 --> 01:10:13.080
have already signed their contracts and
are already working in a new

01:10:13.080 --> 01:10:17.080
technologies unit in a special employment
center and with which it has had

01:10:17.080 --> 01:10:19.920
an impact on their life, on their life
project and the improvement of their

01:10:19.920 --> 01:10:21.360
future.

01:10:22.800 --> 01:10:26.800
And before I say goodbye, as I'm not sure
how I'm doing with time, I think I'm

01:10:26.800 --> 01:10:30.800
doing alright, I'll leave you with these
recommendations that were very useful

01:10:30.800 --> 01:10:34.360
to me back in the day, which you'll
probably already know, but take a

01:10:34.360 --> 01:10:37.360
look at the INTEF website of the
Ministry, it's fantastic.

01:10:37.440 --> 01:10:38.560
There are so many things.

01:10:38.560 --> 01:10:42.920
The Educational Technology Observatory
program we created is wonderful.

01:10:43.080 --> 01:10:45.720
It opened up a whole new world for us
when we had no idea what to do, and you

01:10:45.720 --> 01:10:47.080
also

01:10:47.080 --> 01:10:50.760
have a fantastic course on artificial
intelligence and education, full

01:10:50.760 --> 01:10:52.640
inclusion which is also essential.

01:10:52.640 --> 01:10:53.600
As of today.

01:10:53.600 --> 01:10:57.600
They have many digital education and
digital skills manuals

01:10:57.600 --> 01:11:00.920
for people with intellectual
disabilities.

01:11:01.360 --> 01:11:05.360
The ICT educational resources you see
there are directly from the CREA

01:11:05.360 --> 01:11:09.360
website that I mentioned earlier, which
focuses on diversity in

01:11:09.360 --> 01:11:12.880
Navarre, and there are many applications
and recommendations.

01:11:13.360 --> 01:11:17.360
The board games that I think we are going
to turn the AI ​​week of our center

01:11:17.360 --> 01:11:20.480
into board games, which allows you to
work on all the computational thinking

01:11:20.480 --> 01:11:22.080
skills.

01:11:22.080 --> 01:11:26.080
And it is true that there is a new ICAS,
which is the digital

01:11:26.080 --> 01:11:30.080
section of the department, and they are
putting a lot of

01:11:30.080 --> 01:11:34.080
effort into taking away some of the
technological aspect.

01:11:34.080 --> 01:11:38.080
Then you have there, and in the end,
there are international

01:11:38.080 --> 01:11:42.080
teaching entities and other organizations
that also have

01:11:42.080 --> 01:11:46.080
a lot of material there, from directly
structured AI programs,

01:11:46.080 --> 01:11:51.720
super simple ones, from ethics, and the
truth is, many very interesting things.

01:11:51.720 --> 01:11:54.000
So that's it, I'm done with this and
that's all.

01:11:54.000 --> 01:11:56.200
I hope it has motivated you.

01:11:56.200 --> 01:11:59.800
Inspired to at least encourage you to
keep doing things in the classroom.

01:11:59.840 --> 01:12:01.080
So, thank you very much.

01:12:04.920 --> 01:12:05.240
Very good.

01:12:05.240 --> 01:12:08.240
After Natalia, it's Chema's turn.

01:12:08.280 --> 01:12:10.560
Then she keeps going up. Then.

01:12:10.560 --> 01:12:13.000
I don't know if it's José María or José
Manuel. I've never known.

01:12:13.000 --> 01:12:17.000
José María González Morón, from the CEIP
Gonzalo

01:12:17.000 --> 01:12:21.000
Fernández de Córdoba, will tell us about
his experience

01:12:21.000 --> 01:12:25.200
with artificial intelligence and with
boys and girls.

01:12:25.240 --> 01:12:27.360
Very well, thank you for being here.

01:12:27.360 --> 01:12:31.360
Well, that's it. First of all, thanks to
Gregorio, Estefanía,

01:12:31.360 --> 01:12:36.040
and María for the invitation, another
year here at this event.

01:12:36.040 --> 01:12:39.040
Thank you for also giving a voice to the
elementary school students.

01:12:39.040 --> 01:12:42.320
It seems that artificial intelligence is
only for older children and even in

01:12:42.320 --> 01:12:44.000
primary school.

01:12:44.000 --> 01:12:46.640
It's been a while since we've been doing
some interesting things, and especially

01:12:46.640 --> 01:12:48.000
because I

01:12:48.000 --> 01:12:50.400
say that we have to educate those
children of the future who are of the

01:12:50.400 --> 01:12:51.640
present, they

01:12:51.640 --> 01:12:54.640
are not of the future, they are of the
present and we have to bring that.

01:12:54.640 --> 01:12:58.640
So I'm going to tell you a little bit
about how we started working a little

01:12:58.640 --> 01:13:02.640
bit with artificial entities at our
school, how we got started, and I'll

01:13:02.640 --> 01:13:06.640
tell you that we started working exactly
in 1920 through the Computational

01:13:06.640 --> 01:13:10.720
Thinking and Artificial Intelligence
School organized by INTEF.

01:13:10.720 --> 01:13:14.720
It was an online course that teachers
from all over Spain signed

01:13:14.720 --> 01:13:18.720
up for, and I was lucky enough to be a
tutor. It changed because the

01:13:18.720 --> 01:13:22.880
previous year it was a school of thought
related to mathematics.

01:13:22.880 --> 01:13:24.680
Then it was super interesting.

01:13:24.680 --> 01:13:28.680
In fact, you can see that even the
regulations today allow us to

01:13:28.680 --> 01:13:32.440
work precisely computationally within
mathematics or

01:13:32.440 --> 01:13:35.440
environmental knowledge according to the
law.

01:13:35.760 --> 01:13:39.760
Well, in Madrid, since we have other
areas, things change, but anyway, it was

01:13:39.760 --> 01:13:43.760
there and precisely at this moment, that
I met, had my first contact with

01:13:43.760 --> 01:13:47.760
artificial intelligence, and just like my
colleague said, that artificial

01:13:47.760 --> 01:13:51.760
intelligence brings everything together,
even though when it was me on that

01:13:51.760 --> 01:13:55.760
day, and I wasn't a visionary, I'm sure
many people thought it, I realized that

01:13:55.760 --> 01:13:59.760
artificial intelligence was going to
devour computational thinking, because

01:13:59.760 --> 01:14:02.560
of course, it obviously gathers
everything and takes it and encompasses

01:14:02.560 --> 01:14:03.960
it, right?

01:14:04.200 --> 01:14:07.240
And I thought it was going to take
longer, but I was

01:14:07.240 --> 01:14:10.240
wrong about that, it was much faster than
I thought.

01:14:10.240 --> 01:14:12.840
He's already eaten it, hasn't he?

01:14:12.840 --> 01:14:16.160
It was precisely in that course that we
started working and we started working

01:14:16.160 --> 01:14:17.840
with Machine Learning for Kids.

01:14:17.840 --> 01:14:21.120
From the famous Doyle and we started
working

01:14:21.120 --> 01:14:24.120
with her and she had a number of
problems.

01:14:24.280 --> 01:14:26.920
The first thing we had to do was use
registration to enter it, which is

01:14:26.920 --> 01:14:28.280
already

01:14:28.280 --> 01:14:31.400
causing us to run into problems with the
Data Protection Act.

01:14:31.400 --> 01:14:35.400
And well, at first what we did was create
a teacher account

01:14:35.400 --> 01:14:38.760
and let the students work on your
account.

01:14:38.800 --> 01:14:42.800
They created their models within your
account,

01:14:42.800 --> 01:14:45.960
but in the end there was a data problem.

01:14:45.960 --> 01:14:49.960
In that case, we did intend to do it only
with

01:14:49.960 --> 01:14:53.080
textual data, without images.

01:14:53.080 --> 01:14:56.520
But of course, if the two images are of a
cloud that belonged to

01:14:56.520 --> 01:14:59.520
this gentleman, then it was even more
complicated.

01:14:59.520 --> 01:15:03.520
So, with a series of data, we started to
create the

01:15:03.520 --> 01:15:07.520
first models, always understanding from
the

01:15:07.520 --> 01:15:11.520
beginning how those artificial learning
machines

01:15:11.520 --> 01:15:15.320
of the scissor worked and how to reach
them.

01:15:15.800 --> 01:15:18.760
To think that in the 19th and 20th
centuries, almost all elementary school

01:15:18.760 --> 01:15:21.760
children didn't know what was artificial
as they do today.

01:15:21.760 --> 01:15:26.040
So, well, we tried to get inside the
artificial entity to see how.

01:15:26.080 --> 01:15:28.240
How it works, not how the artificial
scissors learn, how, how they create

01:15:28.240 --> 01:15:29.360
those models.

01:15:29.600 --> 01:15:31.600
Well, like new textures.

01:15:31.600 --> 01:15:35.600
And then the first exercise we had, which
if you know it, the

01:15:35.600 --> 01:15:39.360
typical one of being able to work with
machine learning, connected perfectly

01:15:39.360 --> 01:15:41.240
with Scratch.

01:15:41.240 --> 01:15:45.240
You could create models with Scratch that
instead of working with the tedious

01:15:45.240 --> 01:15:48.400
Scratch lists, which sometimes, because
of something that

01:15:48.400 --> 01:15:51.400
failed, that is, didn't work, however, it
did work.

01:15:51.400 --> 01:15:55.400
Then we can create the typical action of
turning a light on or

01:15:55.400 --> 01:15:59.400
off, or turning the fan on or off, even
if you had a typo or if there

01:15:59.400 --> 01:16:03.280
was a fan with a V, it worked perfectly
because the artificial

01:16:03.280 --> 01:16:06.280
scissors were behind it and allowed it to
work.

01:16:06.280 --> 01:16:10.280
A list of classes wouldn't have worked
then. Well, we learned those

01:16:10.280 --> 01:16:14.280
first models and well, that, that was
working and then in my school

01:16:14.280 --> 01:16:18.280
we have many projects at a somewhat
different level, as I was saying,

01:16:18.280 --> 01:16:22.280
by the way I have to talk to him and my
namesake, my son Manuel, No,

01:16:22.280 --> 01:16:26.280
those projects interest me, they are
going to university because

01:16:26.280 --> 01:16:30.080
we do a lot for this society with
artificial scissors.

01:16:30.840 --> 01:16:34.840
I'm talking about this project that was
related to the

01:16:34.840 --> 01:16:38.840
topics of sports, physics, and the girls,
boys, and children

01:16:38.840 --> 01:16:42.680
of this team created a project to
encourage people to do

01:16:42.680 --> 01:16:45.680
sports using artificial intelligence.

01:16:45.880 --> 01:16:48.280
Okay, I'm going to show the short video
they created themselves explaining their

01:16:48.280 --> 01:16:49.520
own project,

01:16:49.520 --> 01:16:52.520
already using artificial intelligence and
using machine learning.

01:16:52.760 --> 01:16:55.760
for Kids.

01:17:09.960 --> 01:17:13.560
We did some research and discovered that
the residents of our neighborhood are

01:17:13.560 --> 01:17:15.400
exercising less and less.

01:17:16.160 --> 01:17:20.160
Therefore, we have been talking to
doctors and they have told

01:17:20.160 --> 01:17:24.160
us that the main sedentary diseases are
osteoarthritis,

01:17:24.160 --> 01:17:27.240
obesity, high blood pressure, and
diabetes.

01:17:27.760 --> 01:17:31.760
We want to end diseases so that adults
can

01:17:31.760 --> 01:17:35.760
look healthy, and for this we have
created a

01:17:35.760 --> 01:17:39.760
mobile application called Escape de Body

01:17:39.760 --> 01:17:42.920
which will be based on an escape room.

01:17:43.880 --> 01:17:47.880
Speaking with elite athletes, they told
us that

01:17:47.880 --> 01:17:50.960
the appropriate amount of daily exercise

01:17:50.960 --> 01:17:53.960
time is between 30 and 45 minutes.

01:17:54.360 --> 01:17:58.360
First, to design our skateboard, we
looked at

01:17:58.360 --> 01:18:02.360
our neighborhood and chose a series of
areas to

01:18:02.360 --> 01:18:06.360
create our script. But there's one
special area

01:18:06.360 --> 01:18:09.680
that you don't need to go anywhere to do.

01:18:09.680 --> 01:18:11.600
And this is a Scratch program.

01:18:11.600 --> 01:18:16.280
We did this mainly on a website called
Machine Learning for Kids.

01:18:16.600 --> 01:18:20.520
From there we taught our computers so
that if you sent it a photo, it

01:18:20.520 --> 01:18:23.520
could distinguish which steps it most
resembled.

01:18:23.880 --> 01:18:27.880
Then we put all this into a Scratch toy
and what you

01:18:27.880 --> 01:18:31.880
have to do is make sure it can see from
head to toe and

01:18:31.880 --> 01:18:36.800
there's a little figure and you have to
follow the same steps as it.

01:18:37.120 --> 01:18:41.240
And it depends on how fast you do it and
how safe it is.

01:18:41.240 --> 01:18:45.240
You're referring to that step, that's the
score it will give you.

01:18:45.800 --> 01:18:48.880
We have done 11 tests, each lasting

01:18:48.880 --> 01:18:51.880
approximately three minutes.

01:18:52.160 --> 01:18:53.200
For our tests.

01:18:53.200 --> 01:18:55.240
We have used street furniture.

01:18:55.240 --> 01:18:59.240
For example, one of the tests is an
elliptical trainer, located on Cine

01:18:59.240 --> 01:19:02.520
Street; we have installed an elliptical
trainer there.

01:19:02.560 --> 01:19:06.560
Another test is Escaleras, which is
located on Avenida

01:19:06.560 --> 01:19:12.160
de los Poblados and they will have to go
up and down the stairs for three minutes.

01:19:12.160 --> 01:19:17.440
We spent some time investigating and
found the TNN project of six Network.

01:19:17.440 --> 01:19:20.160
Juan Félix Mateos taught us this.

01:19:20.160 --> 01:19:24.160
We had thought that we could use it for
our application

01:19:24.160 --> 01:19:27.400
because thanks to a button that we will
put in all

01:19:27.400 --> 01:19:30.400
the tests, we will be able to locate the
users.

01:19:30.680 --> 01:19:33.840
We will use a network called Lora.

01:19:34.280 --> 01:19:38.280
This network has very good coverage in
our neighborhood and in our

01:19:38.280 --> 01:19:42.080
school, which we use for a project called
Birds in the Cloud with

01:19:42.080 --> 01:19:45.080
a sensor nest to see if there are any
birds in it.

01:19:45.160 --> 01:19:48.640
We have created a prototype of our
application in an application called App

01:19:48.640 --> 01:19:50.400
Pop.

01:19:50.600 --> 01:19:54.880
You will be able to see this design at
the end of our presentation.

01:19:55.000 --> 01:19:59.000
Each time a user takes tests, points will
be added,

01:19:59.000 --> 01:20:02.560
and those points will go towards a
ranking.

01:20:02.560 --> 01:20:06.560
What will we do to win a prize for our
application?

01:20:06.560 --> 01:20:10.560
We will create a monthly ranking for the
first

01:20:10.560 --> 01:20:14.560
prize, which will be a free voucher that
could

01:20:14.560 --> 01:20:18.920
be used for skating in the south on
occasion, etc.

01:20:20.400 --> 01:20:23.400
Well, this is the project they came up
with.

01:20:23.560 --> 01:20:28.600
Did you know they explored the topic of
artificial intelligence through sports?

01:20:28.600 --> 01:20:30.480
There is a small part, the other part is
also social.

01:20:30.480 --> 01:20:34.480
It's a project that we end up working on
throughout the whole course, and in the

01:20:34.480 --> 01:20:38.480
end there are more things besides the
artificial scissors, but in the end it's

01:20:38.480 --> 01:20:41.480
part of that project and that's where
she's starting to use it.

01:20:41.480 --> 01:20:44.200
And you see that we obviously started in
those courses because of the pandemic and

01:20:44.200 --> 01:20:45.560
the mask.

01:20:45.560 --> 01:20:48.560
It is clear that we started working on it
in that course.

01:20:48.560 --> 01:20:52.560
Well, right after the pandemic, just like
the colleague

01:20:52.560 --> 01:20:56.320
who mentioned Code, the first artificial
intelligence

01:20:56.320 --> 01:20:59.320
courses started to appear at Code.

01:20:59.680 --> 01:21:03.680
Here you can see some coding through the
artificial

01:21:03.680 --> 01:21:08.640
license, the impact of artificial,
generative energy and so on.

01:21:08.640 --> 01:21:12.320
But how does AI work specifically for me?

01:21:12.760 --> 01:21:13.600
Everything is a little bit more.

01:21:13.600 --> 01:21:18.000
The one I really liked was the theme of
and for the oceans.

01:21:18.040 --> 01:21:22.040
I think the S medium for primary school
is fundamental for them to understand

01:21:22.040 --> 01:21:26.040
how, how artificial intelligence learns
from models, how, how it makes

01:21:26.040 --> 01:21:29.480
mistakes, because it doesn't have, hasn't
trained well or doesn't have enough

01:21:29.480 --> 01:21:31.240
models, right?

01:21:31.240 --> 01:21:35.240
So, through that program that I
recommend, reading

01:21:35.240 --> 01:21:39.520
across the oceans, they learn how it
works from the inside.

01:21:39.520 --> 01:21:42.840
AI is like adding little machine things
and then

01:21:42.840 --> 01:21:45.840
we see this code stuff, come on, let's
try it.

01:21:45.840 --> 01:21:48.120
So above all we're getting ahead of
ourselves with sounds.

01:21:48.120 --> 01:21:52.120
And having seen this, well, with this it
is possible, in 5th grade,

01:21:52.120 --> 01:21:55.960
in 4th grade they can start working and
they begin to understand how AI works

01:21:55.960 --> 01:21:57.920
from the inside.

01:21:58.240 --> 01:22:02.240
I don't know if you've seen Code yet, but
when

01:22:02.240 --> 01:22:06.240
Code.org started with the famous "Jesus"

01:22:06.240 --> 01:22:09.880
from Obama, we're going to learn by 2025.

01:22:09.880 --> 01:22:12.840
I think it was 2000, so I don't know how
many programmers in the world have

01:22:12.840 --> 01:22:14.320
changed completely.

01:22:14.320 --> 01:22:16.960
And now, as you can see, AI is in the
world, and now the goal is no longer for

01:22:16.960 --> 01:22:18.280
children to learn

01:22:18.280 --> 01:22:21.280
to program, but for children to learn
artificial intelligence.

01:22:21.280 --> 01:22:25.280
Most courses are now focused, still in
contact, but almost all courses

01:22:25.280 --> 01:22:28.280
are now focused on children learning
artificially.

01:22:28.280 --> 01:22:32.280
I also want to, I also want to launch the
ESA because we always say

01:22:32.280 --> 01:22:36.280
that children are not digital natives and
so on, and I think that

01:22:36.280 --> 01:22:40.280
children are current and so it is
artificial to use it in the same

01:22:40.280 --> 01:22:44.280
way that any new little doll comes out in
the bazaar and the next

01:22:44.280 --> 01:22:48.280
day it is in school and they teach you
what that little doll is that

01:22:48.280 --> 01:22:51.560
you don't know because you don't know
what it is.

01:22:51.640 --> 01:22:52.800
Well, in this there is an equal.

01:22:52.800 --> 01:22:57.320
Ultimately, they turn to AI just like we
turn to anything else.

01:22:57.480 --> 01:23:00.920
Are they there, are they in the world, so
they are boys and girls

01:23:00.920 --> 01:23:03.920
of today and they arrive and they know
him, right?

01:23:04.240 --> 01:23:08.240
Well, at school, after this whole spiel I
just gave, we met

01:23:08.240 --> 01:23:11.520
Learning Ml and Juanda and this changed
everything.

01:23:11.560 --> 01:23:12.720
It changed us too. Why?

01:23:12.720 --> 01:23:16.480
Because Machine Learning for Kids we had
those problems with the male, all the

01:23:16.480 --> 01:23:18.400
data, and the registration.

01:23:18.440 --> 01:23:22.440
And of course, ml already offered us
another way of seeing things,

01:23:22.440 --> 01:23:26.400
because it was open, we could use it,
there was no need to register.

01:23:26.440 --> 01:23:30.240
You could even download it, have it on
your desktop, and work from

01:23:30.240 --> 01:23:33.240
your model without having to send the
data anywhere.

01:23:33.240 --> 01:23:37.880
Of course, that's ultimately beneficial
for an educational institution, isn't it?

01:23:38.040 --> 01:23:42.560
Yes, I imagine you already know about it,
that you've seen the snow.

01:23:42.880 --> 01:23:44.760
I don't have to explain anything about
what this is.

01:23:44.760 --> 01:23:47.280
You know him perfectly well. But well, it
was a whole world.

01:23:47.280 --> 01:23:48.720
In fact, it didn't look like that
anymore.

01:23:48.720 --> 01:23:49.520
It looks much nicer now.

01:23:49.520 --> 01:23:54.200
This image is of the previous model; it
looks much, much nicer now.

01:23:54.200 --> 01:23:58.320
Then we started using Learning ml at
school and

01:23:59.280 --> 01:24:01.040
Well,

01:24:01.040 --> 01:24:04.640
No, I haven't mentioned it, but we've
been working on the topic

01:24:04.640 --> 01:24:07.640
of technology, robotics, and so on for a
long time.

01:24:07.720 --> 01:24:11.720
So from third grade onwards, everything
good that even from first

01:24:11.720 --> 01:24:15.720
grade onwards, but right now, but from
scratch, everything in the

01:24:15.720 --> 01:24:19.720
class does a project that we then present
at a science fair that will

01:24:19.720 --> 01:24:23.720
be in 15 days at school, that in May we
do it at the end of May in a science

01:24:23.720 --> 01:24:28.440
fair and everyone presents their small
project or their big project.

01:24:28.800 --> 01:24:31.680
There are some that have to do with
exercise, some that are useful to us.

01:24:31.680 --> 01:24:35.680
One of the ones that came out was this
one that

01:24:35.680 --> 01:24:39.240
I know was for a kind of underwater robot

01:24:39.240 --> 01:24:42.240
that recycled the seas, right?

01:24:42.600 --> 01:24:47.280
If you look at the model, they
differentiate between marine animals.

01:24:47.280 --> 01:24:50.880
Because of course, if his robot
accidentally picked up a marine animal,

01:24:50.880 --> 01:24:52.680
it had to return it to the sea.

01:24:52.680 --> 01:24:56.240
And then there was the waste that they
saw was most

01:24:56.240 --> 01:24:59.240
abundant in the sea: plastic, rubber, and
so on.

01:25:00.680 --> 01:25:03.160
We took photographs with children.

01:25:03.160 --> 01:25:07.160
Well, they collected quite a bit of data,
but they also took photographs

01:25:07.160 --> 01:25:12.400
of small plastic models and things that
came into the classroom to handle, right?

01:25:13.640 --> 01:25:16.520
Here is the plastic part too, which is
also very important.

01:25:16.520 --> 01:25:20.520
They made their own seabed with

01:25:20.520 --> 01:25:24.520
origami and then on top what they did

01:25:24.520 --> 01:25:29.760
here with the Lego pieces was their own
machine.

01:25:30.000 --> 01:25:34.000
But all of that was directed with a
machine from China that Jorge Lobo

01:25:34.000 --> 01:25:38.000
will surely talk about a lot when he
comes, and we programmed it as machine

01:25:38.000 --> 01:25:42.640
learning because you can connect machine
learning to China and work with it.

01:25:42.640 --> 01:25:46.640
So, in order to identify a marine animal
made

01:25:46.640 --> 01:25:50.640
with origami or rubber plastic, the robot
that

01:25:50.640 --> 01:25:54.640
you will see here in the next image, you
see that

01:25:54.640 --> 01:25:58.640
it was here, will allow us to introduce
waste

01:25:58.640 --> 01:26:02.640
into its corresponding space and if not,
the

01:26:02.640 --> 01:26:06.640
marine animal was just on the left which
has a

01:26:06.640 --> 01:26:09.960
small hole and would return to the
seabed.

01:26:09.960 --> 01:26:12.960
But they tell it better than I can.

01:26:13.040 --> 01:26:16.040
I'm going to show you the video and you
can watch it.

01:26:28.600 --> 01:26:31.960
We started this super challenge by

01:26:31.960 --> 01:26:34.960
researching fish extinction.

01:26:37.240 --> 01:26:41.240
And we decided that we wanted to end

01:26:41.240 --> 01:26:45.920
ocean pollution, so we started our model.

01:26:45.920 --> 01:26:49.920
We watched some tutorials for

01:26:49.920 --> 01:26:53.480
making origami sea animals.

01:26:56.680 --> 01:27:00.600
Next, we began to paint the seabed.

01:27:02.520 --> 01:27:06.520
This is what our submarine will be doing,
which

01:27:06.520 --> 01:27:10.520
will take the garbage to our sorter; the
only

01:27:10.520 --> 01:27:15.560
thing we were missing to complete our
project were the areas.

01:27:18.320 --> 01:27:22.320
When we finished the crafting part, we
started to

01:27:22.320 --> 01:27:27.360
undertake Artificial Intelligence, the
basis of our invention.

01:27:29.280 --> 01:27:32.280
This is our finished project.

01:27:32.920 --> 01:27:36.920
This is our finished submarine

01:27:36.920 --> 01:27:41.080
and this is our waste sorting depot.

01:27:42.640 --> 01:27:46.640
The submarine goes underwater and then
the robot

01:27:46.640 --> 01:27:50.640
collects the waste, which then goes to
our sorter,

01:27:50.640 --> 01:27:53.960
where we classify all the waste.

01:27:55.320 --> 01:27:58.480
This is how the program that recognizes
waste

01:27:58.480 --> 01:28:01.480
with artificial intelligence works.

01:28:01.520 --> 01:28:05.520
And now my classmate showed a bottle and
recognized

01:28:05.520 --> 01:28:10.880
that it is plastic, and now she showed a
fish and recognized that too.

01:28:14.080 --> 01:28:18.080
Then, using the Chinese board and

01:28:18.080 --> 01:28:23.960
Scratch, we managed to get our robot's
waste sorted.

01:28:24.880 --> 01:28:27.960
As you can see, it's a marine animal and
we're returning it.

01:28:28.000 --> 01:28:31.000
the sea.

01:28:31.840 --> 01:28:35.160
As you can see here, my colleague is
showing a scratch wheel.

01:28:35.160 --> 01:28:36.040
Do you recognize it?

01:28:36.040 --> 01:28:38.480
We threw it in the right place.

01:28:48.960 --> 01:28:51.680
And she runs with the bottle.

01:28:51.680 --> 01:28:55.680
As you can see, the waste is perfectly

01:28:55.680 --> 01:28:59.120
sorted and the sea is cleaner

01:28:59.120 --> 01:29:02.120
and less polluted.

01:29:07.680 --> 01:29:11.240
Well, as you can see, it's another
project that was already in place, that

01:29:11.240 --> 01:29:13.040
was done with machine learning.

01:29:13.080 --> 01:29:17.800
And well, each project has its own theme
and all that.

01:29:18.160 --> 01:29:22.160
Well, right after that whole pandemic,

01:29:22.160 --> 01:29:25.520
regulations governing AI also appear.

01:29:25.960 --> 01:29:29.960
I don't know if you're familiar with it,
but we in Madrid are

01:29:29.960 --> 01:29:33.960
fortunate to have Decree 61 2022, which
is the curriculum

01:29:33.960 --> 01:29:37.400
for the Primary Education stage, where
schools have

01:29:37.400 --> 01:29:40.400
the option to include the technology
area.

01:29:40.400 --> 01:29:43.400
robotics.

01:29:44.480 --> 01:29:49.160
Just like the second foreign language in
all or one of them.

01:29:49.720 --> 01:29:52.360
In primary education courses, or in any
course we want, we can really include

01:29:52.360 --> 01:29:53.720
this area

01:29:53.720 --> 01:29:57.160
from first grade, 6th grade, or in all
courses, or in any course we want.

01:29:57.160 --> 01:30:01.160
This area can be included in this, it
says that it responds to a series

01:30:01.160 --> 01:30:05.160
of societal needs and has three important
areas: electricity,

01:30:05.160 --> 01:30:09.160
the mechanics of computational
experimentation, and other

01:30:09.160 --> 01:30:13.160
disciplines, one of which is artificial
intelligence, which is

01:30:13.160 --> 01:30:16.160
also addressed, in addition to the Y, O,
and others.

01:30:16.200 --> 01:30:18.840
Well, look at 2022, we already have
something in Madrid that talks about

01:30:18.840 --> 01:30:20.200
artificial

01:30:20.200 --> 01:30:22.840
intelligence in primary education, about
how to start working on it, what that

01:30:22.840 --> 01:30:24.200
decree

01:30:24.200 --> 01:30:26.840
asks of us or what it says we have to
work on. Well, look at knowing and

01:30:26.840 --> 01:30:28.200
valuing the

01:30:28.200 --> 01:30:31.760
possibilities that artificial
intelligence has to act with machines or

01:30:31.760 --> 01:30:34.760
systems in a way that facilitates work
and gradually improves it.

01:30:34.760 --> 01:30:38.760
What he's really asking us to do is allow
children to learn about these virtual

01:30:38.760 --> 01:30:42.640
social intelligence machines from the
inside and how they work, right?

01:30:42.640 --> 01:30:45.280
Apart from that, it's about seeing if you
can see the evolution of people, knowing

01:30:45.280 --> 01:30:46.640
about computer

01:30:46.640 --> 01:30:49.280
applications or emerging digital
technologies or artificial antigens,

01:30:49.280 --> 01:30:50.640
integrating programs

01:30:50.640 --> 01:30:54.160
into machines or systems that allow their
autonomy in a responsible way.

01:30:54.160 --> 01:30:58.160
So you're asking us to use machine
learning and

01:30:58.160 --> 01:31:01.280
then, for concerts and skills, to

01:31:01.280 --> 01:31:04.280
learn the principles of using AI.

01:31:04.280 --> 01:31:08.280
There he addressed the issue of ethics,
biases, all the

01:31:08.280 --> 01:31:12.280
things that I think children can perhaps
understand in a

01:31:12.280 --> 01:31:16.280
simple way, that there are people behind
AI and I think

01:31:16.280 --> 01:31:20.280
they should know this because they use it
freely and without

01:31:20.280 --> 01:31:25.840
anyone watching them and they need to
know that there is something behind it.

01:31:25.960 --> 01:31:28.880
No, things aren't always free, are they?

01:31:30.000 --> 01:31:32.680
Then the strategy of the first automatic.

01:31:32.680 --> 01:31:35.200
How they learn, how they learn, how they
learn from their data.

01:31:35.200 --> 01:31:39.200
And that's why I also think it's
important that we learn from

01:31:39.200 --> 01:31:43.040
these systems so that they input data and
see how the data

01:31:43.040 --> 01:31:46.040
they don't provide is generated, right?

01:31:46.040 --> 01:31:48.080
And then, well, that's it.

01:31:48.080 --> 01:31:52.080
The series of applications for
Introduction to Artificial Agency,

01:31:52.080 --> 01:31:57.240
or putting something that we do, was
meant for more complicated topics, right?

01:31:57.240 --> 01:32:00.600
They don't have to learn basic things
about how AI works.

01:32:00.640 --> 01:32:03.080
If you notice, it doesn't talk about
generative AI either.

01:32:03.080 --> 01:32:07.040
We are talking more about learning models
than generative learning, obviously.

01:32:07.040 --> 01:32:12.640
We'll eventually have to work on it, but
that's not the end of it, is it?

01:32:13.120 --> 01:32:14.440
And what for?

01:32:14.440 --> 01:32:17.400
What else do we use AI for in school?

01:32:17.400 --> 01:32:18.720
Well, in the end, so did we.

01:32:18.720 --> 01:32:21.360
In 5th and 6th grade, with these projects
that we mentioned at the end, we give it

01:32:21.360 --> 01:32:22.720
an approach

01:32:22.720 --> 01:32:25.640
that is why I said that I was very
interested in what José Manuel said about

01:32:25.640 --> 01:32:28.560
the social issues of working, therefore
he works a lot on this.

01:32:28.600 --> 01:32:32.600
We focus heavily on the SDGs and
ultimately we want all the

01:32:32.600 --> 01:32:36.160
5th and 6th grade students to do a team
project that

01:32:36.160 --> 01:32:39.160
has something to do with one of the SDGs.

01:32:39.360 --> 01:32:42.000
In 5th grade we don't force them to do a
technology project, but in 6th grade we

01:32:42.000 --> 01:32:43.360
force them to do a

01:32:43.360 --> 01:32:46.520
technology project, even though sometimes
in 5th grade some students do a

01:32:46.520 --> 01:32:48.520
technology project and projects come out
that are so shady, I think, like this

01:32:48.520 --> 01:32:49.520
one.

01:32:49.520 --> 01:32:53.880
Thinking about disability, you can check
out this really cool project.

01:33:08.600 --> 01:33:12.160
Sometimes we don't realize that there are
people who have more difficulty moving

01:33:12.160 --> 01:33:13.960
around and encounter many obstacles.

01:33:13.960 --> 01:33:16.920
This worries us a lot and we want to help
you.

01:33:16.920 --> 01:33:20.880
We have been investigating and have
concluded that the biggest problems these

01:33:20.880 --> 01:33:23.880
people face are related to mobility and
communication.

01:33:23.880 --> 01:33:27.800
That's why we've created a mobile app
called Disability that will help these

01:33:27.800 --> 01:33:29.800
people with everything they need.

01:33:29.840 --> 01:33:34.640
The application is designed to be highly
accessible for all types of disabilities.

01:33:34.680 --> 01:33:39.200
In our main menu, each person will select
their disability in each of them.

01:33:39.200 --> 01:33:41.320
Receive the necessary help.

01:33:41.320 --> 01:33:45.760
For example, people, as a reduced
measure, will have a kind of Google Maps.

01:33:45.760 --> 01:33:49.760
It will warn them about wider streets
with fewer pedestrians, and even alert

01:33:49.760 --> 01:33:53.680
them to possible architectural barriers,
construction work, or similar issues that

01:33:53.680 --> 01:33:55.680
they should be careful of.

01:33:56.520 --> 01:34:00.520
The app will send alerts just like car
navigation apps do and

01:34:00.520 --> 01:34:04.960
will organize the best routes to get to
the desired destination.

01:34:05.200 --> 01:34:10.280
Now let's see how our app works with a
person with reduced mobility.

01:34:10.280 --> 01:34:15.240
I need to get to the bakery by the
fastest route, please.

01:34:17.280 --> 01:34:18.200
Straight ahead.

01:34:18.200 --> 01:34:20.240
Turn left in 50 meters.

01:34:20.240 --> 01:34:23.240
You'll see.

01:34:29.840 --> 01:34:35.320
Now let's see how the degrees work with a
blind person.

01:34:40.240 --> 01:34:41.880
Where can I go from here?

01:34:41.880 --> 01:34:44.880
Turn 90 degrees to the left.

01:34:45.640 --> 01:34:47.640
Careful, there's a small curb.

01:34:47.640 --> 01:34:50.280
Advance.

01:34:50.280 --> 01:34:51.000
Advance.

01:34:51.000 --> 01:34:54.000
Six meters.

01:34:58.720 --> 01:34:59.360
Careful!

01:34:59.360 --> 01:35:02.360
There's another small curb.

01:35:02.840 --> 01:35:03.960
It has a ramp.

01:35:03.960 --> 01:35:06.920
Advance two meters.

01:35:06.920 --> 01:35:09.160
Turn right.

01:35:09.160 --> 01:35:13.120
Now he has a wide, empty street where he
can walk.

01:35:13.120 --> 01:35:16.120
Don't worry.

01:35:16.320 --> 01:35:18.960
We have also seen that people are
influenced by trends.

01:35:18.960 --> 01:35:22.400
They have a lot of communication problems
when it comes to greeting people.

01:35:27.960 --> 01:35:28.960
I can't help you.

01:35:28.960 --> 01:35:31.960
I don't understand you.

01:35:34.920 --> 01:35:37.920
I don't understand what you're trying to
tell me.

01:35:46.240 --> 01:35:49.440
Our app will translate and interpret your

01:35:49.440 --> 01:35:52.440
signs and allow us to show you the
answers.

01:35:52.440 --> 01:35:55.440
The message of signs.

01:35:56.000 --> 01:35:59.000
We also spoke with funding organizations
within our

01:35:59.000 --> 01:36:02.000
app to raise money for the manufacture of
prosthetics.

01:36:02.040 --> 01:36:04.680
We will donate that money to associations
like Ayúdame 3D with which we

01:36:04.680 --> 01:36:06.040
collaborate,

01:36:06.040 --> 01:36:09.080
which are dedicated to manufacturing
prostheses with 3D companies

01:36:09.080 --> 01:36:12.080
and sending them to countries with few
resources in Africa.

01:36:13.480 --> 01:36:16.000
Our app is now available on the Play
Store.

01:36:16.000 --> 01:36:20.000
Help us spread the word, because with it
we will make

01:36:20.000 --> 01:36:23.320
life a little easier for the people
around us.

01:36:23.680 --> 01:36:27.680
Furthermore, by using the mobility data
from our

01:36:27.680 --> 01:36:30.760
phones, we need many downloads to make

01:36:30.760 --> 01:36:33.760
our system a little more reliable each
day.

01:36:34.480 --> 01:36:37.560
Download disability information and

01:36:37.560 --> 01:36:40.560
collaborate with us to help people in
need.

01:36:44.120 --> 01:36:45.960
Once every time it becomes available.

01:36:45.960 --> 01:36:50.280
The app for the app was done.

01:36:50.520 --> 01:36:54.520
That's how they had done it, 1/1 of
Google's manual, which is very

01:36:54.520 --> 01:36:58.520
complicated, but the language part, if
they had done it, they

01:36:58.520 --> 01:37:02.520
knew, they had done how to translate all
the images into text, the

01:37:02.520 --> 01:37:06.520
part that they had done, that when they
made the video it still

01:37:06.520 --> 01:37:09.520
didn't work, but it did convert the
symbols to text.

01:37:09.520 --> 01:37:11.280
We did manage to do that.

01:37:11.280 --> 01:37:15.280
And then, with these crazy things about
projects, I would

01:37:15.280 --> 01:37:19.280
like to finish by showing this project
that I find very

01:37:19.280 --> 01:37:23.280
interesting, because sometimes when we do
projects at

01:37:23.280 --> 01:37:27.040
school there are themes that make you
say, "I don't know what will come of

01:37:27.040 --> 01:37:28.960
this."

01:37:28.960 --> 01:37:32.600
This group thought about suicide and I
said I don't know what project we're

01:37:32.600 --> 01:37:34.440
going to come up with here.

01:37:34.440 --> 01:37:37.840
So, well, I'm going to show you the video
because something quite interesting came

01:37:37.840 --> 01:37:39.560
out of it.

01:37:40.000 --> 01:37:44.000
We are from [unclear] and we've come to
talk about a topic

01:37:44.000 --> 01:37:48.000
that's been discussed a lot lately. This
topic is suicide,

01:37:48.000 --> 01:37:51.080
and there aren't many solutions to
prevent it.

01:37:51.200 --> 01:37:54.200
One of the solutions is the telephone of
hope.

01:37:54.520 --> 01:37:58.520
If a suicidal person calls this number,
they will receive help from a

01:37:58.520 --> 01:38:02.760
psychologist, but with these calls we try
to prevent suicide in a friendly way.

01:38:04.160 --> 01:38:08.800
We have realized that in Spain there were
4.3 suicides.

01:38:09.280 --> 01:38:10.000
At least.

01:38:10.000 --> 01:38:14.000
Spain is at level 59 on the suicide list
of all

01:38:14.000 --> 01:38:17.480
countries, with Greenland in first place.

01:38:19.040 --> 01:38:22.240
And let's hope these high numbers come
down.

01:38:23.160 --> 01:38:28.360
Okay, okay, lots of talk about suicides,
but what about our candle project?

01:38:28.400 --> 01:38:32.080
We want suicidal people to chat with
people who share their interests and

01:38:32.080 --> 01:38:33.920
become friends.

01:38:33.920 --> 01:38:36.160
But how will we achieve that?

01:38:36.160 --> 01:38:40.160
First, we investigated and found that
many suicidal people post that they

01:38:40.160 --> 01:38:44.160
want to take their own lives on social
networks like Instagram. We would

01:38:44.160 --> 01:38:48.160
detect certain posts and send a message
to the suicidal person who posted

01:38:48.160 --> 01:38:51.640
that message on social networks. That
message would lead them

01:38:51.640 --> 01:38:54.640
to a questionnaire with questions about
their tastes.

01:38:55.880 --> 01:38:59.880
If we see that another person who has
also taken the questionnaire has similar

01:38:59.880 --> 01:39:03.880
tastes to the other person, we would put
them both in a private chat so they

01:39:03.880 --> 01:39:07.080
can talk to each other and get to know
each other that way.

01:39:07.080 --> 01:39:09.320
They knew, by word of mouth and with
luck, through vicarious means, about the

01:39:09.320 --> 01:39:10.440
two suicides.

01:39:11.680 --> 01:39:16.600
We hope this video has been informative.

01:39:19.120 --> 01:39:23.120
You see, in the end it seems that from
something like this an

01:39:23.120 --> 01:39:27.120
interesting project emerged because they
saw that there was

01:39:27.120 --> 01:39:30.160
a bot that could discover comments within
social

01:39:30.160 --> 01:39:33.160
networks and it was pretty cool, let's
say.

01:39:33.400 --> 01:39:35.480
And now that we're working on this, I'm
going to give you a scoop.

01:39:35.480 --> 01:39:38.600
We will be presenting this soon at the
Federation and at the fair we have this

01:39:38.600 --> 01:39:40.200
June.

01:39:40.200 --> 01:39:44.200
But look, now we're working on a
recycler, but this year, this

01:39:44.200 --> 01:39:48.200
time with a conveyor belt that then has
that container on the

01:39:48.200 --> 01:39:52.920
right that rotates and things fall into
it. It was fully automated.

01:39:53.120 --> 01:39:56.480
The one in the fish tank was manual.

01:39:57.080 --> 01:39:59.200
We're going to do everything this way.
We're still working on it.

01:39:59.200 --> 01:40:02.280
Today we changed the recycling bin
because it had gotten quite heavy.

01:40:02.280 --> 01:40:04.840
I think we were talking about the
cardboard final, but well, we're getting

01:40:04.840 --> 01:40:06.160
there and they're working on it.

01:40:06.160 --> 01:40:09.200
Same with China, Ml with and with the
Chinese plate.

01:40:09.200 --> 01:40:12.520
We are working with her to be able to
connect to all of this.

01:40:13.080 --> 01:40:17.080
And then we're also working on a really
cool process, which you'll

01:40:17.080 --> 01:40:21.080
know as Microbit AI Create, which is for
working with the Microbit

01:40:21.080 --> 01:40:25.080
board, especially in movement projects
like the one we saw at the

01:40:25.080 --> 01:40:29.080
beginning. Now you can work on them
because they put the Microbit

01:40:29.080 --> 01:40:33.080
on your wrist or leg and you can control
a series of movements your

01:40:33.080 --> 01:40:38.320
body makes and see if they do them well
or not, or do more or less exercise.

01:40:38.520 --> 01:40:41.760
And there's a really cool thing we're
working on, which is an anti-gender

01:40:41.760 --> 01:40:43.400
violence button.

01:40:43.440 --> 01:40:47.440
So, based on a woman's movements, she can
have the

01:40:47.440 --> 01:40:50.440
bracelet, the microbit, and how does

01:40:50.440 --> 01:40:53.440
he know which pattern made the bracelet?

01:40:53.480 --> 01:40:57.240
Well, she could be calling an emergency
number to report that there is

01:40:57.240 --> 01:40:59.120
gender-based violence.

01:40:59.120 --> 01:41:01.680
So I say that really cool projects come
out of them, things that seem like

01:41:01.680 --> 01:41:04.280
nothing, but really cool little things
come out of them.

01:41:04.760 --> 01:41:05.320
And nothing.

01:41:05.320 --> 01:41:08.480
For my part, I thanked everyone for
listening and I hope you liked the

01:41:08.480 --> 01:41:10.080
projects.

01:41:15.720 --> 01:41:18.720
We have time for.

01:41:18.840 --> 01:41:19.160
Very good.

01:41:19.160 --> 01:41:19.880
Thank you very much, Chema.

01:41:19.880 --> 01:41:22.760
We have time for a few questions.

01:41:22.760 --> 01:41:27.040
Look here in the aisle to my right, in
the 5th row.

01:41:27.760 --> 01:41:30.640
Maria.

01:41:30.640 --> 01:41:34.640
I teach elementary school, and sometimes
I find it a bit overwhelming how to reach

01:41:34.640 --> 01:41:37.360
elementary school students, and I've
loved all the projects you've done.

01:41:37.360 --> 01:41:38.760
Congratulations!

01:41:38.760 --> 01:41:42.760
The thing is, it makes me a little
doubtful,

01:41:42.760 --> 01:41:46.760
with all the limitations that the
Community

01:41:46.760 --> 01:41:50.000
of Madrid is putting on us regarding

01:41:50.000 --> 01:41:53.000
the use of computers and screens.

01:41:53.040 --> 01:41:54.160
How do you do it?

01:41:54.160 --> 01:41:58.160
Because what it makes me feel is that
what they're doing is slowing down

01:41:58.160 --> 01:42:01.560
all these projects, meaning we can only
use it once a week.

01:42:02.120 --> 01:42:05.960
What they're doing to me is telling me to
just leave it, because we

01:42:05.960 --> 01:42:08.960
don't give you the opportunity to do all
these things.

01:42:08.960 --> 01:42:12.960
So, despite all these limitations, you
manage to carry

01:42:12.960 --> 01:42:16.960
out these projects. The robotics
technology area,

01:42:16.960 --> 01:42:20.960
because it's outside of that, allows you
to have an

01:42:20.960 --> 01:42:24.320
additional session to work on this within
the area.

01:42:24.320 --> 01:42:27.120
We are outside of that regulation, the
one that allows us to work in session,

01:42:27.120 --> 01:42:29.920
apart from what is done in class, and
well, then regulate the rest.

01:42:29.920 --> 01:42:34.040
But hey, with that I now work as a team,
sharing devices and so on.

01:42:34.080 --> 01:42:37.440
In the end they do come out eventually,
but basically from the

01:42:37.440 --> 01:42:40.440
area of ​​robotics technology, basically
from there.

01:42:40.800 --> 01:42:44.800
How many hours do you have in a session
per week? In the session

01:42:44.800 --> 01:42:48.800
per week itself, we have two, because of
the same project, but

01:42:48.800 --> 01:42:52.800
well, one session per week is... but we
start from first grade,

01:42:52.800 --> 01:42:56.200
we imprison them from first grade until
6th grade.

01:42:56.200 --> 01:42:59.880
We have everything there, including many
emotional prevention programs;

01:42:59.880 --> 01:43:02.880
it's all a bit lower, and what we do is
gradually creating.

01:43:02.920 --> 01:43:06.280
We were teaching about artificial
intelligence, 3D design, and other

01:43:06.280 --> 01:43:08.000
things.

01:43:08.000 --> 01:43:12.000
By the time they get to this point, we'll
show you the artificial project,

01:43:12.000 --> 01:43:16.000
but in the virtual project, 3D printing
projects appear, projects from

01:43:16.000 --> 01:43:19.840
each person appear, it comes from the
theme, from what they have been working

01:43:19.840 --> 01:43:21.800
on, we appear there.

01:43:21.840 --> 01:43:25.840
I am a big supporter of using technology,
but of creating things with technology,

01:43:25.840 --> 01:43:28.720
because I believe that they should not be
so much users but creators with

01:43:28.720 --> 01:43:30.160
technology.

01:43:30.240 --> 01:43:33.320
So, that's the idea that they end up
having a

01:43:33.320 --> 01:43:36.320
project that ultimately changes things.

01:43:36.480 --> 01:43:39.160
You're going to talk to him here, but
when we do the science fair at school,

01:43:39.160 --> 01:43:40.520
you'll see him on the table.

01:43:41.000 --> 01:43:44.240
When children talk about one of their

01:43:44.240 --> 01:43:47.240
projects, they speak in a different way.

01:43:47.240 --> 01:43:49.200
They bring it to you, they take it there
to the project and they make you fall in

01:43:49.200 --> 01:43:50.200
love with it.

01:43:50.200 --> 01:43:54.200
He knows what's good, but with the area,
I'm telling you,

01:43:54.200 --> 01:43:58.200
the area is fundamental for this,
especially because,

01:43:58.200 --> 01:44:03.760
look, you already include it in the
resume and you have a guaranteed time.

01:44:03.760 --> 01:44:07.760
You work on it during the week, but when
it's not within an area, it's

01:44:07.760 --> 01:44:11.760
because the math teacher, whoever wants
to work on it, or doesn't want

01:44:11.760 --> 01:44:15.760
to work on it, so it can also happen that
there are some courses that are

01:44:15.760 --> 01:44:19.760
worked on, others that are not, so
because the teacher likes it more,

01:44:19.760 --> 01:44:23.400
maybe less, because of the robotics
technology. I go back to my

01:44:23.400 --> 01:44:26.400
school, but we make sure that everyone
goes through it.

01:44:26.400 --> 01:44:29.400
I think it's something, something good
for that reason.

01:44:31.520 --> 01:44:32.560
Well, welcome back.

01:44:32.560 --> 01:44:37.880
After the break, I have Jorge Lobo here
who is going to speak to us.

01:44:37.880 --> 01:44:41.880
It's also primary school and she's going
to talk to us now, we're going

01:44:41.880 --> 01:44:45.800
to go off-topic a bit because we're going
to introduce robotics and circuit boards,

01:44:45.800 --> 01:44:47.760
which is this board.

01:44:47.760 --> 01:44:51.720
It's enchanted, I think they keep it in
the community, it's another one.

01:44:52.440 --> 01:44:56.440
The important thing is to grasp the idea
and then transfer

01:44:56.440 --> 01:45:00.400
that knowledge to the plates they give
you in the Community

01:45:00.400 --> 01:45:03.400
of Madrid, which are very similar in
reality.

01:45:03.920 --> 01:45:05.320
That's clear.

01:45:05.320 --> 01:45:08.680
Jorge got it from China because he

01:45:08.680 --> 01:45:11.680
collaborated on the design of that plate.

01:45:11.720 --> 01:45:15.280
So, robotics and artificial intelligence
with e-learning like us, well, there's

01:45:15.280 --> 01:45:17.080
Jorge Lobo.

01:45:17.080 --> 01:45:18.560
Thank you, thank you.

01:45:21.400 --> 01:45:25.400
Well, good afternoon and first of all,

01:45:25.400 --> 01:45:28.960
thank you for being here again.

01:45:29.760 --> 01:45:33.760
And indeed I'm going to talk a little bit
about

01:45:33.760 --> 01:45:37.760
another layer of programming and
artificial

01:45:37.760 --> 01:45:41.800
intelligence, and that's adding that
robotics layer.

01:45:43.160 --> 01:45:46.600
What I'm going to do is talk first about
tools,

01:45:46.600 --> 01:45:49.600
the ones I normally use at school.

01:45:50.640 --> 01:45:54.640
Then I'll show some projects we've
carried out.

01:45:54.640 --> 01:45:58.560
And finally, I'll answer a question that
often

01:45:58.560 --> 01:46:01.560
comes up: how difficult is this?

01:46:03.440 --> 01:46:07.440
Normally, as Gregorio says, I use the
boards

01:46:07.440 --> 01:46:11.440
from China because, among other things,
we

01:46:11.440 --> 01:46:15.440
designed them together: three colleagues,

01:46:15.440 --> 01:46:19.440
José Pujol, who is from secondary school;

01:46:19.440 --> 01:46:23.440
Xavier Rosas, who is from vocational
training;

01:46:23.440 --> 01:46:27.440
and myself, who is from primary school.
We put

01:46:27.440 --> 01:46:31.440
everything we thought was interesting for

01:46:31.440 --> 01:46:35.880
working on robotics and programming into
this board.

01:46:36.640 --> 01:46:40.640
And as for software we have it, which is
a

01:46:40.640 --> 01:46:44.640
clone of Scratch in which you can also

01:46:44.640 --> 01:46:49.640
work with machine learning thanks to
Learning Ml.

01:46:51.800 --> 01:46:55.240
What does the plate have?

01:46:55.240 --> 01:46:59.240
Well, a lot of sensors and actuators,

01:46:59.240 --> 01:47:02.360
which allows us to work from

01:47:02.360 --> 01:47:05.360
primary to secondary.

01:47:05.400 --> 01:47:08.600
I can tell you that my colleagues have
used it in secondary

01:47:08.600 --> 01:47:11.600
school and in vocational training
courses.

01:47:11.640 --> 01:47:15.640
So we have, from a different LED diodes,
RGB

01:47:15.640 --> 01:47:19.640
LEDs, we have a light sensor,
pushbuttons,

01:47:19.640 --> 01:47:23.640
an accelerometer that allows us to know
what

01:47:23.640 --> 01:47:26.800
position the board is in so that by

01:47:26.800 --> 01:47:29.800
moving it, we can also work with it.

01:47:29.840 --> 01:47:33.840
Microphone, that is, a lot of devices,
with

01:47:33.840 --> 01:47:37.840
the advantage that they are already
integrated

01:47:37.840 --> 01:47:41.840
into a board, which is one of the
problems we

01:47:41.840 --> 01:47:47.200
saw mainly when we worked, which I
started quite a while ago.

01:47:47.200 --> 01:47:50.520
I think it was the year 2013 or 2014.

01:47:50.840 --> 01:47:54.840
I worked with Arduino and protoboards,
which are

01:47:54.840 --> 01:47:58.840
boards with holes where you put the wires
and build

01:47:58.840 --> 01:48:02.840
the circuit there, which is very
interesting

01:48:02.840 --> 01:48:06.840
because it allows you to focus on doing
that

01:48:06.840 --> 01:48:10.680
precision work, requiring a lot of
concentration

01:48:10.680 --> 01:48:13.680
to know where you have to connect.

01:48:13.840 --> 01:48:17.840
But it's true that assembling the circuit
took

01:48:17.840 --> 01:48:21.840
quite a bit of time, and with this board,
which was

01:48:21.840 --> 01:48:25.840
initially a shield for Arduino—that is,
it was

01:48:25.840 --> 01:48:29.840
mounted on top of the Arduino board—what
we wanted

01:48:29.840 --> 01:48:35.160
was to focus on programming and leave the
assembly aside a bit.

01:48:35.160 --> 01:48:38.160
the circuit.

01:48:40.000 --> 01:48:42.280
Regarding the software issue.

01:48:42.280 --> 01:48:45.520
I always like to start like this.

01:48:45.520 --> 01:48:49.520
When I talk about artificial intelligence
in the classroom,

01:48:49.520 --> 01:48:53.520
two questions before we start working: Is
it relevant to

01:48:53.520 --> 01:48:57.520
work with artificial intelligence in the
classroom? And

01:48:57.520 --> 01:49:01.400
is it appropriate when we are going to
use a tool?

01:49:01.400 --> 01:49:05.400
I think it's important to be clear that
this tool

01:49:05.400 --> 01:49:09.400
has a justification, a justification that
helps

01:49:09.400 --> 01:49:13.400
us to work on the content we want to
develop and to

01:49:13.400 --> 01:49:18.160
work on those skills we want our students
to acquire.

01:49:18.480 --> 01:49:22.520
This is what I mean by the question of
whether it is relevant.

01:49:22.880 --> 01:49:25.880
And then, on the other hand, if it is
appropriate.

01:49:26.520 --> 01:49:30.520
Although there are some tools that can be

01:49:30.520 --> 01:49:34.520
very interesting for working in the
classroom,

01:49:34.520 --> 01:49:38.520
we also have to keep in mind the ethical

01:49:38.520 --> 01:49:42.520
guidelines on the use of artificial

01:49:42.520 --> 01:49:46.520
intelligence because a tool may be very

01:49:46.520 --> 01:49:52.400
interesting but not comply with, for
example, the data protection we need.

01:49:53.040 --> 01:49:57.040
So it always seems strange that he comes
up with

01:49:57.040 --> 01:50:01.000
this in a talk about artificial
intelligence,

01:50:01.000 --> 01:50:04.000
which is counterintuitive.

01:50:04.000 --> 01:50:07.160
It's not good, but you're going to talk
to us about artificial intelligence, not

01:50:07.160 --> 01:50:08.760
about not using it.

01:50:09.160 --> 01:50:11.360
Nothing could be further from my
intention.

01:50:11.360 --> 01:50:15.360
I am very much in favor of using it, but
also taking

01:50:15.360 --> 01:50:19.360
into account these two aspects, not using
it just for

01:50:19.360 --> 01:50:23.360
the sake of using it, because it is true
that in recent

01:50:23.360 --> 01:50:27.360
years there has been an expansion and
perhaps we

01:50:27.360 --> 01:50:31.360
sometimes let ourselves be blinded by the
glitter

01:50:31.360 --> 01:50:35.160
and the sparkles that these applications
give us.

01:50:35.600 --> 01:50:39.280
And sometimes there are things we can
achieve to work on those

01:50:39.280 --> 01:50:42.280
skills that don't require artificial
intelligence.

01:50:42.280 --> 01:50:45.920
And sometimes there are tools that
perhaps we shouldn't use because they

01:50:45.920 --> 01:50:47.760
don't meet the conditions.

01:50:48.360 --> 01:50:52.360
Since you'll be doing the presentations
later,

01:50:52.360 --> 01:50:56.360
I'm going to skip this a bit quickly to
get to

01:50:56.360 --> 01:51:00.360
the projects, but we need to keep in mind
all

01:51:00.360 --> 01:51:05.320
these features of the applications we're
going to use.

01:51:06.080 --> 01:51:10.080
And one of the first documents that

01:51:10.080 --> 01:51:14.080
discussed the appropriateness of using

01:51:14.080 --> 01:51:18.080
these applications in the classroom is

01:51:18.080 --> 01:51:23.480
the Beijing Consensus, which I also have
linked here.

01:51:23.480 --> 01:51:27.480
A brief summary in case it's of interest,
but basically what

01:51:27.480 --> 01:51:31.480
it tells us is that AI in the teaching
and learning process

01:51:31.480 --> 01:51:35.880
should promote a fair and inclusive use
of artificial intelligence.

01:51:35.920 --> 01:51:39.520
Ensure that artificial intelligence
promotes gender equity and contributes to

01:51:39.520 --> 01:51:41.320
gender equality.

01:51:41.360 --> 01:51:45.120
To ensure the ethical, transparent, and
verifiable application of AI in the

01:51:45.120 --> 01:51:47.040
educational context.

01:51:47.080 --> 01:51:51.080
This got cut off when I made the PDF, but
I imagine

01:51:51.080 --> 01:51:55.080
it will come out fine later and I can do
ongoing

01:51:55.080 --> 01:51:59.160
monitoring, evaluation, and research on
the matter.

01:52:00.560 --> 01:52:04.560
And well, when it comes to discussing

01:52:04.560 --> 01:52:08.560
the suitability of a tool, I find a

01:52:08.560 --> 01:52:12.560
document from INTEF in 2024 very

01:52:12.560 --> 01:52:16.560
interesting, which you can also link

01:52:16.560 --> 01:52:20.560
to here, that tells us how appropriate

01:52:20.560 --> 01:52:24.560
it is to use a tool or not depending on

01:52:24.560 --> 01:52:28.480
the perspective of data protection.

01:52:28.680 --> 01:52:34.360
Here's a screenshot so you can see a
little bit about how he organizes it.

01:52:34.400 --> 01:52:38.400
The tool's name is included, along with a
brief description,

01:52:38.400 --> 01:52:42.400
whether it's free or proprietary
software, if registration

01:52:42.400 --> 01:52:46.280
is required, a series of notes, and,
well, a classification

01:52:46.280 --> 01:52:49.280
that, as you can see, is very intuitive.

01:52:50.280 --> 01:52:54.280
If it's green, then it can be used with
confidence;

01:52:54.280 --> 01:52:58.760
if it's red, then its use is not
recommended.

01:52:58.840 --> 01:53:02.840
If it's yellow, then you need to take

01:53:02.840 --> 01:53:08.280
certain precautions or consider some
factors.

01:53:09.360 --> 01:53:13.360
Having said that, I'm not going to talk
much about

01:53:13.360 --> 01:53:17.360
Learning ml because it has already been
mentioned

01:53:17.360 --> 01:53:21.360
quite a bit and you know it perfectly
well, but it

01:53:21.360 --> 01:53:25.360
is the tool that I have chosen for many
years to work

01:53:25.360 --> 01:53:29.880
with artificial intelligence in the
classroom.

01:53:30.160 --> 01:53:33.440
Just like I mentioned before.

01:53:33.440 --> 01:53:37.440
Chema: Well, I started with Machine
Learning

01:53:37.440 --> 01:53:41.440
for Kids. We had certain problems that he
has

01:53:41.440 --> 01:53:45.440
already mentioned, and at one point
Juanda

01:53:45.440 --> 01:53:49.440
Rodríguez said, "Hey, I can create a tool
that

01:53:49.440 --> 01:53:52.640
can solve these problems," and the

01:53:52.640 --> 01:53:55.640
reasons for working with this tool.

01:53:55.640 --> 01:53:58.840
Learning Ml is the first thing that is
based on free

01:53:58.840 --> 01:54:01.840
software, which is a choice that
interests me a lot.

01:54:02.560 --> 01:54:06.560
In fact, now when trying to put up the
presentation, it

01:54:06.560 --> 01:54:10.560
took me longer than anyone else, because
since it was

01:54:10.560 --> 01:54:14.000
a Windows computer I wasn't familiar with
it.

01:54:14.400 --> 01:54:18.160
I usually use free software for
everything, so

01:54:18.160 --> 01:54:21.160
that's a very important personal choice.

01:54:21.160 --> 01:54:25.160
It can be used without registration, it
does not track or use user data,

01:54:25.160 --> 01:54:28.400
and it has a desktop version here in the
Community of Madrid.

01:54:28.400 --> 01:54:31.200
It is integrated within Max.

01:54:31.200 --> 01:54:35.200
Is the operating system that is
distributed

01:54:35.200 --> 01:54:39.200
free of charge and is open source
software,

01:54:39.200 --> 01:54:43.200
and has a Scratch-based environment, an

01:54:43.200 --> 01:54:47.200
advancement or just another step forward

01:54:47.200 --> 01:54:51.720
when we have worked with programming in
Scratch?

01:54:52.000 --> 01:54:54.120
It's simply a matter of adding a few more
blocks.

01:54:54.120 --> 01:54:57.240
The students somehow feel

01:54:57.240 --> 01:55:00.240
in a safe, familiar place.

01:55:00.280 --> 01:55:04.280
He already knows how the whole workflow
works and

01:55:04.280 --> 01:55:08.280
it's simply a matter of taking one more
step to

01:55:08.280 --> 01:55:11.400
integrate these machine learning blocks.

01:55:11.480 --> 01:55:15.480
And we are fortunate that there is
software available

01:55:15.480 --> 01:55:19.480
for these homemade boards that already
integrates

01:55:19.480 --> 01:55:23.480
Learning ml, so we have a set that allows
us to work on

01:55:23.480 --> 01:55:27.160
programming, robotics and artificial
intelligence.

01:55:27.920 --> 01:55:31.920
The two legs it has are, on the one hand,
Hechiza

01:55:31.920 --> 01:55:35.920
Blocks, which would be all the hardware
control

01:55:35.920 --> 01:55:39.920
part of these boards, and Learning ml, as
you

01:55:39.920 --> 01:55:45.320
know it, is totally integrated within the
software.

01:55:45.320 --> 01:55:49.320
So the Learning ml part is exactly the

01:55:49.320 --> 01:55:53.320
same, but when we click on the y Blocks

01:55:53.320 --> 01:55:57.320
it takes us to this Scratch clone just

01:55:57.320 --> 01:56:01.320
like the Learning ml one, but with

01:56:01.320 --> 01:56:05.320
specific blocks that allow us to work

01:56:05.320 --> 01:56:09.320
with LED diodes, with the RGB LED,

01:56:09.320 --> 01:56:13.320
with the temperature sensor, the MK

01:56:13.320 --> 01:56:17.320
input light sensor, MK, which has the

01:56:17.320 --> 01:56:22.640
same functionality as the Mackey boards.

01:56:22.880 --> 01:56:26.880
In short, we have a fairly comprehensive
system

01:56:26.880 --> 01:56:30.880
on a single board and with a fairly
integrated

01:56:30.880 --> 01:56:34.880
ecosystem where we can work on these
three things:

01:56:34.880 --> 01:56:39.760
robotics programming and artificial
intelligence.

01:56:39.760 --> 01:56:43.760
After this, I'm going to show you some of

01:56:43.760 --> 01:56:47.120
the projects developed at school.

01:56:47.120 --> 01:56:51.120
I haven't said it, right now I'm in the
Code

01:56:51.120 --> 01:56:55.120
School 4.0 program, but my school is Vega
in

01:56:55.120 --> 01:57:00.520
Carabanchel and these are some projects
we've done there.

01:57:02.240 --> 01:57:06.360
This is a

01:57:06.360 --> 01:57:10.360
A slightly older version of something

01:57:10.360 --> 01:57:14.080
Chema has already shown you.

01:57:15.520 --> 01:57:17.600
Let's see if

01:57:17.600 --> 01:57:18.840
This is over.

01:57:18.840 --> 01:57:20.920
The announcement.

01:57:20.920 --> 01:57:24.920
This was the first version of a garbage

01:57:24.920 --> 01:57:28.920
sorter like the one Chema showed, but
with

01:57:28.920 --> 01:57:32.920
a much older version and with a bucket
made

01:57:32.920 --> 01:57:36.920
from recycled materials, with a simple

01:57:36.920 --> 01:57:40.920
cardboard bread box that has a servomotor

01:57:40.920 --> 01:57:44.920
that moves from side to side to be able
to

01:57:44.920 --> 01:57:48.040
sort different types of waste.

01:57:50.240 --> 01:57:53.240
Remove it.

01:57:55.200 --> 01:57:59.200
This is one of the projects I have
repeated

01:57:59.200 --> 01:58:03.200
the most, because it has something that

01:58:03.200 --> 01:58:07.200
interests me a lot, and that is the

01:58:07.200 --> 01:58:13.120
connection of the area of ​​Plastic Arts
with technology.

01:58:13.120 --> 01:58:17.120
Many times, and Chema has mentioned it
before,

01:58:17.120 --> 01:58:21.120
it's very interesting to have the
technology

01:58:21.120 --> 01:58:25.120
area because you have a time dedicated to

01:58:25.120 --> 01:58:29.120
programming, robotics, very specific

01:58:29.120 --> 01:58:33.480
things, but I also find it very
interesting.

01:58:33.880 --> 01:58:37.880
When you apply that robotics, that
artificial

01:58:37.880 --> 01:58:41.880
intelligence, or that programming within
other

01:58:41.880 --> 01:58:45.880
areas, and the area of ​​plastic arts is
fantastic for

01:58:45.880 --> 01:58:49.400
this, because you can work on many
projects that you then animate in some

01:58:49.400 --> 01:58:51.200
way.

01:58:51.240 --> 01:58:55.240
In this case, what we did was create
models

01:58:55.240 --> 01:58:59.240
of some rooms, and if you remember the
room

01:58:59.240 --> 01:59:03.240
they showed us before, in which it said

01:59:03.240 --> 01:59:07.240
"turn on the fan" or "I'm hot" or "I
can't

01:59:07.240 --> 01:59:11.240
see," and through a text model, it would

01:59:11.240 --> 01:59:16.240
turn the light on in Scratch or turn it
off or whatever.

01:59:16.280 --> 01:59:20.280
So we bring that into the physical world

01:59:20.280 --> 01:59:24.280
through this platform where we have an

01:59:24.280 --> 01:59:29.160
assistant that says, "What can I do for
you?"

01:59:29.200 --> 01:59:30.120
I'm hot, okay?

01:59:30.120 --> 01:59:34.400
I turn on the fan and we have the
students' creations.

01:59:34.400 --> 01:59:38.400
As you can see, its different rooms, each

01:59:38.400 --> 01:59:42.400
with its own style, in which a servomotor

01:59:42.400 --> 01:59:46.400
becomes the fan and an LED diode becomes

01:59:46.400 --> 01:59:51.520
the little light that will illuminate the
room.

01:59:52.440 --> 01:59:56.280
This allows each person to express
themselves

01:59:56.280 --> 01:59:59.280
artistically by creating the room as they
wish.

01:59:59.760 --> 02:00:03.760
On one hand, we have the drawing phase,
so to

02:00:03.760 --> 02:00:07.040
speak, in which he decorates his room.

02:00:07.040 --> 02:00:10.760
We have another phase in which, as you
can see, he creates

02:00:10.760 --> 02:00:13.760
3D furniture with cardboard or cardstock.

02:00:13.800 --> 02:00:17.520
I'm going to stop this for a moment
because I also find it interesting.

02:00:18.480 --> 02:00:23.160
If you notice, we have a double room with
two floors here.

02:00:23.160 --> 02:00:23.600
Because?

02:00:23.600 --> 02:00:27.600
Well, because this particular
student—this was a

02:00:27.600 --> 02:00:31.360
student—found it interesting that her
house had two floors.

02:00:31.400 --> 02:00:35.400
In other words, we also have the
possibility of

02:00:35.400 --> 02:00:39.040
expanding, of personalizing, so that each

02:00:39.040 --> 02:00:42.040
person can do it exactly as they like.

02:00:42.120 --> 02:00:46.120
We're going to look at another example
where

02:00:46.120 --> 02:00:50.120
instead of having two floors, it has two
rooms

02:00:50.120 --> 02:00:55.400
but they are next to each other, I think
it's the following one.

02:00:57.240 --> 02:01:01.240
Well, here you can see that what also

02:01:01.240 --> 02:01:05.240
allows us to work with these technologies

02:01:05.240 --> 02:01:08.400
is that the expansion adaptations

02:01:08.400 --> 02:01:11.400
are very simple.

02:01:11.440 --> 02:01:15.440
As you saw earlier, we had a project with
Solo

02:01:15.440 --> 02:01:19.200
el gatito, who was our assistant, and
this

02:01:19.200 --> 02:01:22.200
team consists of two people.

02:01:22.240 --> 02:01:26.240
Since it ended sooner, she decided she
could put a background

02:01:26.240 --> 02:01:31.000
in the room, change some things, and
decorate it differently.

02:01:31.160 --> 02:01:36.720
The expansion comes about in a super
simple way.

02:01:36.760 --> 02:01:38.960
Have you already finished the Minimum
Viable Project?

02:01:38.960 --> 02:01:42.960
Well, now you can improve it by adding
backgrounds,

02:01:42.960 --> 02:01:46.920
adding sounds, adding whatever you want.

02:02:01.600 --> 02:02:04.280
It looks pretty good here.

02:02:04.280 --> 02:02:08.280
These are some small supports we

02:02:08.280 --> 02:02:12.280
made with the 3D printer to hold the

02:02:12.280 --> 02:02:17.200
servomotors, and this is where the LED
went.

02:02:17.200 --> 02:02:19.520
All of this was an improvement.

02:02:19.520 --> 02:02:23.520
This was the second year we did it
because in the

02:02:23.520 --> 02:02:27.520
first year they were inserted into the
room

02:02:27.520 --> 02:02:31.520
itself and then with practice you realize
that

02:02:31.520 --> 02:02:35.520
you can make it simpler, in this case
with the

02:02:35.520 --> 02:02:38.840
supports to simply fit them into the
room.

02:02:38.880 --> 02:02:42.880
You save time, you also save them

02:02:42.880 --> 02:02:46.880
from having to make cuts and from

02:02:46.880 --> 02:02:51.400
having to somehow break into their rooms.

02:02:51.440 --> 02:02:55.120
And well, these are small improvements
that are

02:02:55.120 --> 02:02:58.120
also made with the different iterations.

02:02:58.160 --> 02:03:01.160
Here you can see what I was telling you
about.

02:03:03.120 --> 02:03:07.120
Another of the most representative rooms,
which in

02:03:07.120 --> 02:03:11.040
this case were two, one next to the other
and which had a small door to connect

02:03:11.040 --> 02:03:13.040
them.

02:03:13.040 --> 02:03:17.040
Anyway, what I was saying are

02:03:17.040 --> 02:03:22.440
things that are quite customizable.

02:03:23.600 --> 02:03:26.960
This is another project.

02:03:26.960 --> 02:03:30.960
This was taken to the Science

02:03:30.960 --> 02:03:34.960
Fair last year and what we have

02:03:34.960 --> 02:03:40.160
is a model image in which again the.

02:03:40.480 --> 02:03:45.720
Art education, the visual arts, has a lot
to do with it.

02:03:45.760 --> 02:03:49.040
Here we have two little animals, a dog
and a cat.

02:03:49.080 --> 02:03:51.520
This is a practice they did in pairs.

02:03:51.520 --> 02:03:55.520
Each pair agreed on who would play the
dog,

02:03:55.520 --> 02:03:59.520
who would play the cat, and then a model
was

02:03:59.520 --> 02:04:03.520
trained with food that was either for
dogs

02:04:03.520 --> 02:04:08.440
or cats, showing the food in question to
the camera.

02:04:08.800 --> 02:04:13.680
The dog was happy wagging its tail, or
the cat was wagging its tail.

02:04:15.320 --> 02:04:18.320
And here we have it.

02:04:27.840 --> 02:04:31.080
Presenting their work at the Science
Fair.

02:04:31.400 --> 02:04:35.400
And the interesting thing about this, as

02:04:35.400 --> 02:04:39.400
you can imagine, is not only doing it,
but

02:04:39.400 --> 02:04:43.400
having these meeting points where
students

02:04:43.400 --> 02:04:46.960
present their work and expose it to

02:04:46.960 --> 02:04:49.960
people, in this case adults.

02:04:49.960 --> 02:04:53.960
And it's fantastic because it's such a
personal

02:04:53.960 --> 02:04:57.960
project, and it exudes such confidence;
it's

02:04:57.960 --> 02:05:03.280
something they've worked hard on, and
it's definitely worth seeing.

02:05:03.280 --> 02:05:07.280
Another thing we've been doing at school
for a while

02:05:07.280 --> 02:05:11.280
now—I mention this because it might be
useful as an

02:05:11.280 --> 02:05:15.280
idea—is putting on small exhibitions for
families

02:05:15.280 --> 02:05:19.280
with all these things, not just science
and technology

02:05:19.280 --> 02:05:23.280
stuff, but also content they've worked on
during the

02:05:23.280 --> 02:05:26.520
course, instead of doing the typical

02:05:26.520 --> 02:05:29.520
festival we've done for many years.

02:05:29.520 --> 02:05:33.520
What we do is exhibitions of their
classwork,

02:05:33.520 --> 02:05:37.520
and I think it's really worthwhile
because

02:05:37.520 --> 02:05:41.160
they are exhibiting work they have

02:05:41.160 --> 02:05:44.160
invested a lot of time in.

02:05:44.200 --> 02:05:49.000
Furthermore, in our case, the students
choose what they want to present.

02:05:49.640 --> 02:05:53.640
In fact, there are some years when they
keep technology

02:05:53.640 --> 02:05:57.640
projects that interest me more and that I
think are really

02:05:57.640 --> 02:06:01.680
cool, and they put in others like showing
their notebooks.

02:06:02.120 --> 02:06:06.120
I think it's very interesting to give
them

02:06:06.120 --> 02:06:10.120
these exhibition spaces because it also
tells

02:06:10.120 --> 02:06:14.120
us what their interests are and, as I
said

02:06:14.120 --> 02:06:18.120
before, what their own work is like, the
work

02:06:18.120 --> 02:06:22.120
they've invested time in. They're very
happy

02:06:22.120 --> 02:06:26.120
to exhibit it, and we realize that
they've

02:06:26.120 --> 02:06:30.120
understood things and have been able to

02:06:30.120 --> 02:06:34.120
transmit that understanding to others,
which

02:06:34.120 --> 02:06:38.120
leads us to know that the teaching and
learning

02:06:38.120 --> 02:06:41.120
process has been very fruitful.

02:06:42.880 --> 02:06:46.880
Another project, a parking barrier

02:06:46.880 --> 02:06:50.880
that recognizes license plates, in

02:06:50.880 --> 02:06:54.880
this case we had to change it and instead

02:06:54.880 --> 02:07:00.640
of numbers we changed it to geometric
shapes,

02:07:00.640 --> 02:07:03.640
because the

02:07:04.480 --> 02:07:07.840
The model with numbers wasn't working
quite right.

02:07:08.000 --> 02:07:12.000
And we arrived at this compromise
solution

02:07:12.000 --> 02:07:15.560
which also served us quite well.

02:07:17.720 --> 02:07:21.720
And one of the latest projects

02:07:21.720 --> 02:07:25.720
in which we are also using

02:07:25.720 --> 02:07:28.960
Lego pieces is...

02:07:29.680 --> 02:07:34.040
Look, it's not going full screen for me,
but I think you can see it pretty well.

02:07:34.040 --> 02:07:38.040
It's a box made with Lego pieces that
opens when

02:07:38.040 --> 02:07:41.240
it detects the face it has been trained

02:07:41.240 --> 02:07:44.240
on, and closes if it doesn't.

02:07:48.560 --> 02:07:50.840
And now,

02:07:50.840 --> 02:07:54.160
As I was saying, how difficult is this?

02:07:57.240 --> 02:08:01.880
Well, it's as difficult as we want it to
be.

02:08:01.880 --> 02:08:06.720
On the one hand, what's important is
what's behind these practices.

02:08:06.760 --> 02:08:10.760
In other words, it's very interesting

02:08:10.760 --> 02:08:14.760
that when we do a project, these
artifices

02:08:14.760 --> 02:08:18.760
or artifacts that we create are like the

02:08:18.760 --> 02:08:22.760
culmination. But there is prior work

02:08:22.760 --> 02:08:26.760
where the learning really connects and

02:08:26.760 --> 02:08:31.600
where that process is really taking
place.

02:08:31.840 --> 02:08:35.840
We've seen things related to robotics

02:08:35.840 --> 02:08:39.840
because that's what the presentation
asked

02:08:39.840 --> 02:08:43.840
for, but if we look at other examples,
for

02:08:43.840 --> 02:08:47.840
instance, one of the jobs we usually do
is

02:08:47.840 --> 02:08:51.880
artificial intelligence for the
environment.

02:08:52.520 --> 02:08:53.240
So that?

02:08:53.240 --> 02:08:54.080
For whatever you want.

02:08:54.080 --> 02:08:58.080
They have a research phase to determine
how we can use

02:08:58.080 --> 02:09:01.920
artificial intelligence to help us
improve the environment.

02:09:01.960 --> 02:09:05.960
Something that comes up often because
it's recurring is the

02:09:05.960 --> 02:09:09.960
topic of recycling, but other topics have
also come up, such as

02:09:09.960 --> 02:09:14.560
vehicles, which are the most polluting
and which are the least polluting?

02:09:14.600 --> 02:09:18.440
How can we help reduce the ozone layer,

02:09:18.440 --> 02:09:21.440
i.e., through free projects?

02:09:21.760 --> 02:09:25.760
Through these investigations, they create
their

02:09:25.760 --> 02:09:29.760
learning process and integration with
robotics,

02:09:29.760 --> 02:09:33.760
programming, or artificial intelligence.

02:09:33.800 --> 02:09:35.640
It's like the final period.

02:09:35.640 --> 02:09:39.640
I want to link this because earlier we
were asking, or I was

02:09:39.640 --> 02:09:45.160
asking, a colleague how to combine all
this technology with the new regulations?

02:09:45.200 --> 02:09:49.200
It's true that it doesn't work well
because we do have

02:09:49.200 --> 02:09:54.800
fewer hours of screen time and fewer
hours where we can use technology.

02:09:55.120 --> 02:09:59.120
We're going to have less effective time,
but we also have to

02:09:59.120 --> 02:10:03.960
think that there's an important part that
doesn't have to be with screens.

02:10:04.240 --> 02:10:08.240
When we are doing this type of work there
are some

02:10:08.240 --> 02:10:12.240
preliminary steps of research, of content
creation

02:10:12.240 --> 02:10:16.240
that can be done without screens, even
when we are

02:10:16.240 --> 02:10:20.240
working with Learning ml and we are going
to create

02:10:20.240 --> 02:10:23.880
a text model, an image model in primary
school.

02:10:23.880 --> 02:10:27.560
I don't make numerical models, but that
could also be the case.

02:10:28.120 --> 02:10:32.120
We can even write and search for those

02:10:32.120 --> 02:10:36.120
photographs or make those drawings
beforehand,

02:10:36.120 --> 02:10:40.120
for which we don't need screens, and that
type

02:10:40.120 --> 02:10:43.880
of screen is the effective time in which
we

02:10:43.880 --> 02:10:46.880
finish shaping that project.

02:10:48.520 --> 02:10:50.520
So,

02:10:50.520 --> 02:10:54.520
We can get into all that preliminary work
as much

02:10:54.520 --> 02:10:58.520
as we want, and the part about creating a
model and

02:10:58.520 --> 02:11:01.560
robotizing it is actually quite short,

02:11:01.560 --> 02:11:04.560
which is what I wanted to show you.

02:11:05.360 --> 02:11:07.600
This will happen at some point.

02:11:07.600 --> 02:11:11.600
I've done it live with Programming the

02:11:11.600 --> 02:11:15.880
board and seeing how it's done in five
minutes.

02:11:16.280 --> 02:11:20.280
But since we have the problem here of
what these

02:11:20.280 --> 02:11:24.600
computers and such have to be, well, I'll
take advantage of it

02:11:24.600 --> 02:11:27.680
a video of a

02:11:27.680 --> 02:11:31.680
an event in which I also did it through
videoconference,

02:11:31.680 --> 02:11:35.680
which I'm going to leave for you so we
can see that

02:11:35.680 --> 02:11:39.680
in just five minutes it detects when
there is a

02:11:39.680 --> 02:11:45.160
bird so that we can have it, look how it
can scare it away.

02:11:45.800 --> 02:11:49.800
So we're going to add a class of images

02:11:49.800 --> 02:11:53.800
that will be bird, I'm going to put

02:11:53.800 --> 02:11:57.600
see and another one that will be

02:11:58.080 --> 02:12:01.080
not to see.

02:12:02.640 --> 02:12:05.680
And now, directly with the camera.

02:12:09.240 --> 02:12:13.240
Let's see if it opens for me, and well, I
don't know. If you're connecting

02:12:13.240 --> 02:12:17.680
from your camera, disconnect the camera
and it will definitely work.

02:12:17.680 --> 02:12:18.240
Okay?

02:12:18.240 --> 02:12:19.040
That's what I was going to say.

02:12:19.040 --> 02:12:22.040
I mean, maybe he's fighting with...

02:12:24.520 --> 02:12:26.360
Exactly.

02:12:26.360 --> 02:12:28.160
Okay,

02:12:28.160 --> 02:12:31.200
So here we're going to take a few

02:12:31.200 --> 02:12:34.200
pictures without a bird in this case.

02:12:34.200 --> 02:12:37.760
Well, it'll be me. Let's see, 21 people
came up.

02:12:37.800 --> 02:12:38.960
That's very good.

02:12:38.960 --> 02:12:42.960
And now we're going to take a few

02:12:42.960 --> 02:12:46.960
pictures for when there is a

02:12:46.960 --> 02:12:52.000
bird, which in this case will also be me.

02:12:52.880 --> 02:12:55.880
But with this mask.

02:12:56.120 --> 02:13:02.080
So we're going to take a similar number
of photographs as well.

02:13:05.240 --> 02:13:08.880
And we're going to give him the
opportunity to learn.

02:13:09.160 --> 02:13:13.160
Since it's an image model, it takes a
little longer than if

02:13:13.160 --> 02:13:16.720
it were a text model, but not too much,
as you can see.

02:13:16.840 --> 02:13:20.600
Then the next one is for the class
camera, because you might have

02:13:20.600 --> 02:13:23.600
trouble with the camera. Oh, yes, yes,
yes, thank you.

02:13:24.600 --> 02:13:28.680
The next step will be to test our model.

02:13:29.440 --> 02:13:32.520
We're going to do it through the camera
and

02:13:32.520 --> 02:13:35.520
in this case it says there are no birds.

02:13:36.240 --> 02:13:39.240
What if I had the mask?

02:13:40.000 --> 02:13:42.760
Well, it detects that yes, there is a
bird.

02:13:42.760 --> 02:13:46.760
Once we have the model, I'm going to
directly

02:13:46.760 --> 02:13:49.840
transfer the camera to the camera.

02:13:51.560 --> 02:13:54.040
Okay, thanks.

02:13:54.040 --> 02:13:57.760
The controls appear here and it confuses
me because I can't easily switch between

02:13:57.760 --> 02:13:59.640
them.

02:13:59.640 --> 02:14:03.640
Okay, so I'm going to create a very
simple

02:14:03.640 --> 02:14:09.600
project that will be triggered by
pressing the green flag.

02:14:11.120 --> 02:14:14.120
forever.

02:14:14.320 --> 02:14:18.320
If it detects that there is a bird,

02:14:18.320 --> 02:14:23.360
we will do that with the Learning ml
blocks.

02:14:23.360 --> 02:14:27.360
In this case, when classifying an

02:14:27.360 --> 02:14:31.360
image that we will obtain from the

02:14:31.360 --> 02:14:36.200
video, we will tell it if it classifies
it.

02:14:39.760 --> 02:14:44.280
It is within the category of the bird
category see has.

02:14:44.280 --> 02:14:48.280
It is very important, then, that we use
the same nomenclature,

02:14:48.280 --> 02:14:52.080
the same name that we have given to the
class.

02:14:53.440 --> 02:14:56.680
We are going to propose that it be moved.

02:14:59.000 --> 02:15:02.920
Where are you? Here, the servo

02:15:02.920 --> 02:15:05.920
should go to 35 degrees.

02:15:05.920 --> 02:15:08.600
For example

02:15:08.600 --> 02:15:11.440
for a little while

02:15:11.440 --> 02:15:14.440
0.2 seconds

02:15:15.120 --> 02:15:18.120
and then.

02:15:18.840 --> 02:15:21.880
Go to, for example, 135.

02:15:22.840 --> 02:15:25.640
Okay, so what I'm going to do now is add

02:15:25.640 --> 02:15:28.640
the camera.

02:15:29.480 --> 02:15:32.480
Let's put it on.

02:15:32.800 --> 02:15:35.800
Make it look good.

02:15:40.840 --> 02:15:44.240
And now, yes.

02:15:44.240 --> 02:15:47.680
We can also see

02:15:47.680 --> 02:15:50.680
the scarecrow.

02:15:50.680 --> 02:15:53.680
What if a little bird appears?

02:15:54.840 --> 02:15:59.720
Our scarecrow waves his broom to chase it
away.

02:16:01.000 --> 02:16:05.000
Well, as you can see, all the

02:16:05.000 --> 02:16:09.000
programming of both the model and

02:16:09.000 --> 02:16:14.640
the Scratch program, actually took five
minutes.

02:16:14.680 --> 02:16:18.640
Is it true that with students it's a
little more, but what is important in

02:16:18.640 --> 02:16:20.640
this work?

02:16:20.920 --> 02:16:24.920
That preliminary art work in which they
created

02:16:24.920 --> 02:16:28.920
the scarecrow, which is what ultimately
takes

02:16:28.920 --> 02:16:32.920
the most time in a single session, we can
schedule

02:16:32.920 --> 02:16:36.920
it, but the art sessions, well, they have
been

02:16:36.920 --> 02:16:40.920
two sessions invested in work that
involves

02:16:40.920 --> 02:16:44.920
cutting and pasting, that involves
planning

02:16:44.920 --> 02:16:48.920
the scarecrow and that involves that
design,

02:16:48.920 --> 02:16:52.920
so, well, the effective screen time is
reduced,

02:16:52.920 --> 02:16:56.920
but it is true that there is preliminary
work

02:16:56.920 --> 02:17:00.920
that we can work on in other classes and
in

02:17:00.920 --> 02:17:05.360
primary school, which is where I teach.

02:17:05.600 --> 02:17:09.600
We are fortunate that, since we can
integrate

02:17:09.600 --> 02:17:13.600
everything in a more free way than in

02:17:13.600 --> 02:17:17.600
secondary school, we can make our
adjustments

02:17:17.600 --> 02:17:21.160
and work in a more global way.

02:17:22.240 --> 02:17:24.920
And that's all I wanted to tell you.

02:17:24.920 --> 02:17:28.920
I'm also leaving you a couple of links to
the China website and the Learning Ml

02:17:28.920 --> 02:17:32.400
website, although you've probably already
seen them quite a bit.

02:17:32.600 --> 02:17:35.560
And well, if you have any questions.

02:17:35.560 --> 02:17:37.000
Happy to answer.

02:17:37.000 --> 02:17:40.000
Hello!

02:17:44.360 --> 02:17:46.440
hello, good afternoon.

02:17:46.440 --> 02:17:50.440
I loved the idea of ​​integrating
robotics, programming,

02:17:50.440 --> 02:17:54.440
and artificial intelligence, but I'm
getting the feeling

02:17:54.440 --> 02:17:58.000
that it depends on using this specific
type of board

02:17:58.000 --> 02:18:01.000
from China and in Code School.

02:18:01.000 --> 02:18:04.720
What we have received in the institutes
has been Arduinos and Microbit.

02:18:04.760 --> 02:18:07.760
So.

02:18:07.800 --> 02:18:11.400
I don't know, I'm somewhat related to
what my colleague said earlier, that I'm

02:18:11.400 --> 02:18:13.240
seeing difficulties.

02:18:13.240 --> 02:18:17.080
When it comes to data protection, you
have to be careful and limit sessions

02:18:17.080 --> 02:18:19.000
with screens.

02:18:19.000 --> 02:18:22.520
We need to use a specific type of license
plate that I

02:18:22.520 --> 02:18:25.520
don't know if we were sent with the
school code.

02:18:25.520 --> 02:18:29.000
So, well, this isn't a complaint, it's
just a comment.

02:18:29.000 --> 02:18:33.000
But I'd like to ask if there's a way to
easily integrate

02:18:33.000 --> 02:18:37.360
the kits we have as a school code with
what you're proposing?

02:18:37.560 --> 02:18:39.720
So, the robotics component wouldn't be
integrated?

02:18:39.720 --> 02:18:40.960
How could we do it?

02:18:42.360 --> 02:18:45.360
Obviously,

02:18:45.600 --> 02:18:49.600
The best and most optimized solution

02:18:49.600 --> 02:18:53.600
would be to have the boards, but the

02:18:53.600 --> 02:18:58.360
technology behind the boards is flawed.

02:18:58.400 --> 02:19:01.400
Let's see, can they be programmed the
same way as an Arduino?

02:19:01.800 --> 02:19:05.800
In other words, you can use an Arduino
board,

02:19:05.800 --> 02:19:09.800
you have to install the standard signed
firmware

02:19:09.800 --> 02:19:14.640
that comes directly in the Arduino ID,
you can find it.

02:19:15.160 --> 02:19:19.160
And the problem? The problem, the

02:19:19.160 --> 02:19:23.160
difficulty, is that the software is

02:19:23.160 --> 02:19:27.160
designed to work with the sensors and

02:19:27.160 --> 02:19:31.160
actuators already integrated into the

02:19:31.160 --> 02:19:35.160
board, but you do have the option of
using

02:19:35.160 --> 02:19:38.720
some of the Arduino pins, the inputs and
outputs.

02:19:40.800 --> 02:19:44.600
With the sensors you want and with the
actuators you want.

02:19:44.640 --> 02:19:47.800
In other words, you'll have limited use,

02:19:47.800 --> 02:19:50.800
but you can use it with Arduino boards.

02:19:51.800 --> 02:19:52.480
Okay, thanks.

02:19:52.480 --> 02:19:56.480
And one more question, while we're on the
topic of evaluation,

02:19:56.480 --> 02:20:00.280
could you tell us a little bit about how
you evaluate this?

02:20:00.360 --> 02:20:03.360
Yes, this

02:20:03.440 --> 02:20:07.440
At my school we don't have a technology

02:20:07.440 --> 02:20:11.440
and robotics subject, so I don't have

02:20:11.440 --> 02:20:15.440
these activities here to show you, but

02:20:15.440 --> 02:20:19.440
I usually use an evaluation rubric that

02:20:19.440 --> 02:20:23.440
is combined, it comes with the content

02:20:23.440 --> 02:20:27.440
parts and with the parts, so to speak,

02:20:27.440 --> 02:20:30.640
of programming, robotics or

02:20:30.640 --> 02:20:33.640
artificial intelligence.

02:20:33.680 --> 02:20:39.640
They are very simple rubrics; what I do
is ask questions.

02:20:39.680 --> 02:20:43.680
I say ask because apart from doing them
myself,

02:20:43.680 --> 02:20:49.560
each person self-evaluates with the same
rubric, slightly simplified.

02:20:49.800 --> 02:20:53.800
So, for example, in the one where we

02:20:53.800 --> 02:20:57.800
saw that I really like the rooms, the

02:20:57.800 --> 02:21:01.800
questions, which I also adapt a lot,

02:21:01.800 --> 02:21:05.800
are things like, "The room has turned

02:21:05.800 --> 02:21:09.800
out very well, a little so-so, it

02:21:09.800 --> 02:21:13.800
could be better," that is, with super

02:21:13.800 --> 02:21:17.800
plain and simple language so that

02:21:17.800 --> 02:21:21.080
each student can identify the

02:21:21.080 --> 02:21:24.080
state of their room.

02:21:24.680 --> 02:21:26.400
Of its creation.

02:21:26.400 --> 02:21:30.400
And I related it to a word that there

02:21:30.400 --> 02:21:34.400
are some of, that are very everyday

02:21:34.400 --> 02:21:38.760
and each one has a punctuation.

02:21:38.800 --> 02:21:42.800
Yes, it was the room, it turned out very
well, it's

02:21:42.800 --> 02:21:46.800
three points and it turned out okay, two
points, if

02:21:46.800 --> 02:21:51.360
it can be improved one point, then
another thing that is.

02:21:51.360 --> 02:21:55.360
Regarding the Scratch program with
Learning

02:21:55.360 --> 02:21:59.360
ml, have you included improvements such
as

02:21:59.360 --> 02:22:03.360
audio? I'm speaking from memory and I
don't

02:22:03.360 --> 02:22:07.360
remember exactly, but have you included

02:22:07.360 --> 02:22:12.200
improvements such as audio, other
characters, a series like that?

02:22:12.240 --> 02:22:16.560
Whether quite a few, some, or not, I
haven't done it.

02:22:16.600 --> 02:22:19.800
And then, according to that rubric, they
are

02:22:19.800 --> 02:22:22.800
evaluated one by one in a fairly simple
way.

02:22:23.160 --> 02:22:24.240
Thank you very much. You're welcome.

02:22:27.880 --> 02:22:31.880
All this taking into account that the
evaluation

02:22:31.880 --> 02:22:35.880
is not only of the practice, but of
everything

02:22:35.880 --> 02:22:41.680
integrated, and that these grades, for
example, go towards art education.

02:22:41.680 --> 02:22:44.680
Then.

02:22:48.440 --> 02:22:49.680
Longer.

02:22:49.680 --> 02:22:52.680
We have time for one more question, if
there is one.

02:22:53.320 --> 02:22:56.320
Thank you.

02:23:00.960 --> 02:23:03.960
Thinking skills?

02:23:05.200 --> 02:23:08.200
Can you repeat the question so it's
recorded? Yes.

02:23:08.600 --> 02:23:12.600
Good afternoon, If the rubric you propose

02:23:12.600 --> 02:23:16.600
also assesses higher-order thinking

02:23:16.600 --> 02:23:20.640
skills, the rubrics are very simple.

02:23:21.080 --> 02:23:24.360
The short answer is no,

02:23:24.360 --> 02:23:27.360
because of the rubrics.

02:23:27.360 --> 02:23:31.240
What interests me is that the students

02:23:31.240 --> 02:23:34.240
also evaluate themselves.

02:23:34.480 --> 02:23:36.520
So

02:23:36.520 --> 02:23:40.520
There are some where they do teamwork,

02:23:40.520 --> 02:23:44.520
for example, they do ask questions about

02:23:44.520 --> 02:23:48.520
how they have been in that team, the
level

02:23:48.520 --> 02:23:52.520
of participation, but they are all

02:23:52.520 --> 02:23:56.520
questions very focused on rating or

02:23:56.520 --> 02:24:01.920
evaluating their work from the object
they have created.

02:24:02.880 --> 02:24:05.880
Thank you so much.

02:24:07.960 --> 02:24:09.800
Thank you very much, José.

02:24:09.800 --> 02:24:12.800
Well, that's it.

02:24:13.480 --> 02:24:15.120
We see the thanks to Jorge again.

02:24:15.120 --> 02:24:18.120
Thank you so much.

02:24:22.480 --> 02:24:22.920
Very good.

02:24:22.920 --> 02:24:24.840
And we're going to move on from...

02:24:24.840 --> 02:24:28.680
We've had two presentations for
elementary school and now we're going to

02:24:28.680 --> 02:24:30.640
move on to secondary school.

02:24:31.720 --> 02:24:37.480
Well, first of all, thank you very much
for calling me to present the project.

02:24:38.800 --> 02:24:42.800
I come from a high school, El Burgo
Ignacio,

02:24:42.800 --> 02:24:46.880
which is in Las Rozas and we have
students who are.

02:24:47.160 --> 02:24:48.480
We have this precedent center.

02:24:48.480 --> 02:24:52.480
TGD is also a center for Vocational
Training, Training Cycles,

02:24:52.480 --> 02:24:56.320
Electricity, Electronics, Computing and
Communications.

02:24:56.640 --> 02:24:59.120
And then it also has classrooms of
excellence inside.

02:24:59.120 --> 02:25:03.120
So we have there a mix that is very much
a mix of students, but

02:25:03.120 --> 02:25:06.800
at the same time also for doing
technology projects.

02:25:06.800 --> 02:25:10.800
It's very interesting because we have
very good students, we have

02:25:10.800 --> 02:25:14.800
many students with ASD who may not seem
like it, but when it comes to

02:25:14.800 --> 02:25:18.320
programming, artificial intelligence and
such, it turns out they are especially

02:25:18.320 --> 02:25:20.120
good at it.

02:25:20.160 --> 02:25:24.160
And then, with the issue of vocational
training cycles, we have

02:25:24.160 --> 02:25:28.160
teachers who I think are very technical,
not just in technology,

02:25:28.160 --> 02:25:32.160
since I'm from a tech background, but
there are many teachers of

02:25:32.160 --> 02:25:36.160
electronics, electricity, people who can
help you with many things

02:25:36.160 --> 02:25:39.200
related to technology, so it's very
interesting.

02:25:39.200 --> 02:25:42.520
I really like this center, and in fact I
haven't been here long, but I like it a

02:25:42.520 --> 02:25:44.200
lot for that reason.

02:25:44.200 --> 02:25:48.200
So, at the beginning of the school year I
took over

02:25:48.200 --> 02:25:52.200
4th year of ESO (Compulsory Secondary
Education)

02:25:52.200 --> 02:25:56.200
and well, I love active methodologies
and, well,

02:25:56.200 --> 02:26:00.200
that children do real projects that they
can then

02:26:00.200 --> 02:26:03.680
apply to solving problems in society.

02:26:04.360 --> 02:26:06.600
Then it occurred to us that they
participate in retouching, which I'm sure

02:26:06.600 --> 02:26:07.720
you know about.

02:26:07.720 --> 02:26:11.720
The Endesa Foundation's ES competition

02:26:11.720 --> 02:26:15.720
is a technology competition whose final

02:26:15.720 --> 02:26:19.720
format is a fair, similar to the Science

02:26:19.720 --> 02:26:23.720
Fair mentioned earlier, where students

02:26:23.720 --> 02:26:27.720
present their projects. Teachers from

02:26:27.720 --> 02:26:31.720
different schools and institutes visit,

02:26:31.720 --> 02:26:35.720
and the students have to "sell" their

02:26:35.720 --> 02:26:39.720
product and convince the visiting

02:26:39.720 --> 02:26:43.720
teachers, who will ultimately vote in

02:26:43.720 --> 02:26:47.720
the final round. Different prizes are

02:26:47.720 --> 02:26:51.720
awarded depending on the category, so

02:26:51.720 --> 02:26:55.720
the students have to convince the
teachers,

02:26:55.720 --> 02:26:59.720
present their project, create brochures

02:26:59.720 --> 02:27:02.800
and posters, and explain it well.

02:27:02.800 --> 02:27:06.600
So, we decided to participate in the

02:27:06.600 --> 02:27:09.600
Endesa competition and...

02:27:11.320 --> 02:27:15.320
The idea is to introduce service
learning, which is to do

02:27:15.320 --> 02:27:19.480
technology projects that serve to solve a
social problem.

02:27:19.480 --> 02:27:22.480
Then in.

02:27:23.440 --> 02:27:27.440
So we started thinking, okay, what are we
going to do? The

02:27:27.440 --> 02:27:31.440
current topic is artificial intelligence,
the humanoids

02:27:31.440 --> 02:27:35.440
that are coming very soon, well, they're
already here, but

02:27:35.440 --> 02:27:39.440
it seems they're going to fill society in
a few years, and

02:27:39.440 --> 02:27:43.440
then also, but how could we connect the
humanoid and artificial

02:27:43.440 --> 02:27:47.760
intelligence with what it can contribute
to us and to them?

02:27:48.120 --> 02:27:51.640
Well, in the end they'll tell you because
I've been compiling different videos, but

02:27:51.640 --> 02:27:53.440
I'll give a brief summary.

02:27:54.160 --> 02:27:56.680
So, how could they help?

02:27:56.680 --> 02:27:59.160
How could a humanoid help in a high
school?

02:27:59.160 --> 02:28:02.560
Then the topic of bullying came up, the
topic of school harassment and...

02:28:02.680 --> 02:28:06.120
And not only that, but also the emotions
that a teenager can experience in a high

02:28:06.120 --> 02:28:07.880
school.

02:28:08.040 --> 02:28:12.040
And then the idea that we had, a little
bit

02:28:12.040 --> 02:28:16.040
among the whole class, was to make a
humanoid

02:28:16.040 --> 02:28:19.480
that would serve to accompany the

02:28:19.480 --> 02:28:22.480
secondary school student.

02:28:22.520 --> 02:28:25.400
Well, it could also be useful for
schools,

02:28:25.400 --> 02:28:27.880
so that

02:28:27.880 --> 02:28:31.360
The students could ask him related
questions.

02:28:31.800 --> 02:28:32.720
I'm nervous.

02:28:32.720 --> 02:28:36.240
I don't know why I'm feeling this, and
the

02:28:36.240 --> 02:28:39.240
humanoid will answer you as best as I
can.

02:28:39.280 --> 02:28:40.720
Why are you nervous?

02:28:40.720 --> 02:28:44.040
Because you're not handling this well as

02:28:44.040 --> 02:28:47.040
responses related to their emotions.

02:28:47.600 --> 02:28:48.160
So.

02:28:48.160 --> 02:28:52.160
Well, from there on the topic, first send

02:28:52.160 --> 02:28:56.160
a video of the team and so on, and then

02:28:56.160 --> 02:29:00.160
halfway through they ask you to send a

02:29:00.160 --> 02:29:03.320
video of what was done in the end.

02:29:03.360 --> 02:29:06.440
Well, this is the 16th, the fair is on
June 16th, so we still have some work to

02:29:06.440 --> 02:29:08.000
do.

02:29:08.160 --> 02:29:11.800
But anyway, this is what we've sent

02:29:11.800 --> 02:29:14.800
for now, I hope you like it.

02:29:16.440 --> 02:29:19.440
So, a short video of the idea, right?

02:29:31.440 --> 02:29:34.800
We know that one in three children have
suffered bullying at some

02:29:34.800 --> 02:29:37.800
point in their lives, and one in ten
continues to suffer it.

02:29:37.800 --> 02:29:39.160
Cyberbullying today.

02:29:39.160 --> 02:29:40.360
That's why, of course.

02:29:40.360 --> 02:29:42.680
Etxeberria, a public school in Las Rozas
de Madrid.

02:29:42.680 --> 02:29:44.840
We are very aware of this issue.

02:29:44.840 --> 02:29:48.160
We wanted to combine an artificial
intelligence that is capable of

02:29:48.160 --> 02:29:51.160
helping all students with their
psychological problems.

02:29:51.160 --> 02:29:53.840
Therefore, our project can be directly
linked to a number of sustainable

02:29:53.840 --> 02:29:56.520
development goals, including the 2030
Agenda.

02:29:56.520 --> 02:30:00.520
Well, this project can be applied to
public, semi-private, and private

02:30:00.520 --> 02:30:04.400
schools, because it is a help for all
types of students who need it.

02:30:04.400 --> 02:30:07.560
But it can also be extrapolated to a lot
of companies,

02:30:07.560 --> 02:30:10.560
such as those that need help for their
employees.

02:30:10.840 --> 02:30:12.040
Competitions, challenges.

02:30:12.040 --> 02:30:15.640
We want to participate and it's Ignacio
Echeverría, I'm sure he'll stand out with

02:30:15.640 --> 02:30:17.440
my technological project.

02:30:17.440 --> 02:30:20.840
We want to show that we will overcome any
challenge we face.

02:30:21.080 --> 02:30:25.080
Mate Rap, for the latter, is the math
teacher

02:30:25.080 --> 02:30:29.200
who, on top of everything else, makes
math ugly.

02:30:29.200 --> 02:30:32.520
Then he put it in there. Okay, then.

02:30:32.520 --> 02:30:34.120
Okay, this is the mini video you submit.

02:30:34.120 --> 02:30:37.120
This is the product we are going to
create.

02:30:37.360 --> 02:30:39.720
So

02:30:39.720 --> 02:30:43.280
Then we made it a little longer so that
we haven't delivered this one yet.

02:30:43.280 --> 02:30:47.280
But, to put it simply, more than three
children have suffered bullying at some

02:30:47.280 --> 02:30:50.080
point in their lives, and one in ten
places are currently experiencing

02:30:50.080 --> 02:30:51.480
cyberbullying.

02:30:51.520 --> 02:30:55.040
That's why Echeverría, a public school in
Las Rozas de Madrid, was born.

02:30:55.480 --> 02:30:56.920
Paint them on the mask.

02:30:56.920 --> 02:30:59.920
Excuse me, I've picked the wrong video.

02:30:59.960 --> 02:31:01.160
This one was good.

02:31:01.160 --> 02:31:05.080
Let me give you a brief summary first.
Basically, it's the idea behind the

02:31:05.080 --> 02:31:07.080
project.

02:31:07.360 --> 02:31:11.360
That was it, to generate not only the
guide that spoke, which we do

02:31:11.360 --> 02:31:15.880
with Scratch, or Scratch Stretch, which
is a kind of Scratch but modified.

02:31:17.440 --> 02:31:21.800
But for that, we wanted to put it into
something physical, into a body.

02:31:21.840 --> 02:31:22.640
Okay?

02:31:22.640 --> 02:31:26.600
Then we came up with the idea of
​​projecting it with a projector onto a

02:31:26.600 --> 02:31:28.600
semi-translucent surface.

02:31:28.600 --> 02:31:29.360
Okay.

02:31:29.360 --> 02:31:32.360
So, as a result of that.

02:31:33.040 --> 02:31:37.040
We had to generate the face that they had
already used to

02:31:37.040 --> 02:31:41.040
generate the face to do the programming
in Scratch, which

02:31:41.040 --> 02:31:45.040
meant that they had to generate each
expression of the face

02:31:45.040 --> 02:31:49.040
with it, but then the programming of the
mouth and what it

02:31:49.040 --> 02:31:53.040
says had to be connected with the web
services, these, the

02:31:53.040 --> 02:31:57.040
AWS, the Amazon Web Services, so that it
would be converted

02:31:57.040 --> 02:32:01.040
to speech text and then synchronized with
the mouth and

02:32:01.040 --> 02:32:05.040
those images of the mouth that are
actually the vowels that

02:32:05.040 --> 02:32:09.040
are associated with each syllable, well,
they also had to

02:32:09.040 --> 02:32:12.800
create them and then program them using
lists in Scratch.

02:32:12.920 --> 02:32:18.880
Try associating those syllables you are
saying with the image of the word they.

02:32:18.880 --> 02:32:23.520
He explains it very well there, okay,
now, I'll put it to you later then, okay.

02:32:26.760 --> 02:32:30.320
So they've put AI into the part of the
face,

02:32:30.320 --> 02:32:33.320
that face that it then projects.

02:32:33.360 --> 02:32:37.360
Then they have to introduce something
physical,

02:32:37.360 --> 02:32:41.360
optical, because they put in a projector
and from the

02:32:41.360 --> 02:32:45.360
projector they have to use an angled lens
and from there

02:32:45.360 --> 02:32:49.360
a mirror so that the distance is not so
long and then the.

02:32:50.560 --> 02:32:54.560
Then the AI ​​in the generation of the
face, the programming of the

02:32:54.560 --> 02:32:57.720
face, and then it also has to understand
what you ask it.

02:32:57.720 --> 02:33:02.560
So, to understand what you're asking, we
use Google Teachable Machine.

02:33:02.560 --> 02:33:03.480
As the saying goes, "machine."

02:33:03.480 --> 02:33:07.480
They generate a list of questions that a
teenager can

02:33:07.480 --> 02:33:11.480
ask a kind of psychologist, and then a
battery of

02:33:11.480 --> 02:33:16.640
answers that programs with faces and text
from Amazon Web Services.

02:33:18.160 --> 02:33:22.160
Furthermore, the idea was also like this
concept in psychology

02:33:22.160 --> 02:33:26.160
that one of them came up with, in which
the patient's gestures

02:33:26.160 --> 02:33:31.240
are imitated to make them feel better
when you are talking to them.

02:33:31.240 --> 02:33:35.240
Then they came up with the idea of
​​introducing

02:33:35.240 --> 02:33:39.240
the servos that would nod and shake their

02:33:39.240 --> 02:33:44.400
heads, copying the person who was talking
to him.

02:33:44.800 --> 02:33:48.800
Then they introduced AI with Arduino, and
for

02:33:48.800 --> 02:33:54.160
that I had to give them a kind of recipe
for how to implement it.

02:33:54.200 --> 02:33:58.200
Yes, if you go to Google, the Machine
explains how to integrate

02:33:58.200 --> 02:34:02.200
that Google Matching link, which you can
easily integrate with

02:34:02.200 --> 02:34:06.200
Scratch to connect it to the Arduino
board; you just have to add

02:34:06.200 --> 02:34:10.200
some JavaScript in between. I made them a
sort of guide, but then

02:34:10.200 --> 02:34:14.040
they expand on it a lot because you teach
them a few things and you think, "Wow,

02:34:14.040 --> 02:34:15.960
they learn so much!"

02:34:16.400 --> 02:34:20.480
And we were doing it, and then they did
it much more.

02:34:20.520 --> 02:34:26.240
So, well, I'll show you.

02:34:32.480 --> 02:34:34.640
I'm

02:34:34.640 --> 02:34:38.640
That's it. Well, that's it, the Amazon
Web Services, the machine, the

02:34:38.640 --> 02:34:42.640
topic of physical optics, which I've
already mentioned, Scratch, the

02:34:42.640 --> 02:34:46.640
topic of optics to be able to do that
video mapping of the face on a surface

02:34:46.640 --> 02:34:50.640
and then the creation of the surface was
done with 3D printing, the edge

02:34:50.640 --> 02:34:54.640
of the frame of the face and then we were
working on the plastics syllabus,

02:34:54.640 --> 02:34:58.960
which is also seen in third, in 4th year
plastics are seen in technology.

02:34:58.960 --> 02:35:02.640
So they made it with polypropylene, they
tried to undo the molds, then in the end

02:35:02.640 --> 02:35:04.480
they left it flat.

02:35:04.480 --> 02:35:08.480
Well, in the end we worked on many more
technological things, not

02:35:08.480 --> 02:35:12.040
just artificial intelligence, the neck
motors to feel and

02:35:12.040 --> 02:35:15.040
to deny and then the AI ​​that goes into
the face.

02:35:15.040 --> 02:35:17.400
And those are the basic concepts.

02:35:17.400 --> 02:35:19.520
Then they made this brochure and
everything else too.

02:35:19.520 --> 02:35:22.640
Nothing works with AI.

02:35:27.640 --> 02:35:28.040
Okay.

02:35:28.040 --> 02:35:32.040
So, compiled from the work we've been
doing in class, including

02:35:32.040 --> 02:35:36.040
homemade videos of what we've been doing
with them, I'm showing them

02:35:36.040 --> 02:35:39.760
to you so you can see how it works, and
at the end I'm including two

02:35:39.760 --> 02:35:42.760
more short parts that show a little bit
of how it works.

02:35:42.800 --> 02:35:47.040
We still have a little while until the
16th, but that's it.

02:35:47.040 --> 02:35:48.880
Almost.

02:35:48.880 --> 02:35:52.280
We know that one in three children have
suffered bullying at some

02:35:52.280 --> 02:35:55.280
point in their lives, and one in ten
continues to suffer it.

02:35:55.280 --> 02:35:56.640
Cyberbullying today.

02:35:56.640 --> 02:36:00.200
That's why Echeverría, a public school in
Las Rozas, Madrid, was born from Google.

02:36:00.200 --> 02:36:02.320
We are very aware of this issue.

02:36:02.320 --> 02:36:06.000
We wanted to combine an artificial
intelligence that is capable of

02:36:06.000 --> 02:36:09.000
helping all students with their
psychological problems.

02:36:09.960 --> 02:36:13.960
The project we've created is called
therapy, designed to help

02:36:13.960 --> 02:36:17.960
people who have any difficulties or
problems at school. It

02:36:17.960 --> 02:36:22.720
combines humanoids with artificial
intelligence in a sensitive way.

02:36:23.840 --> 02:36:27.840
Companies like Tesla's Sophia or Optimus
are

02:36:27.840 --> 02:36:31.840
demonstrating that this is the present;
the

02:36:31.840 --> 02:36:35.840
main goal of this project is to develop a
humanoid

02:36:35.840 --> 02:36:39.840
robotic psychologist to help high school

02:36:39.840 --> 02:36:43.840
students with their problems, mainly
bullying,

02:36:43.840 --> 02:36:47.480
thus promoting awareness about bullying.

02:36:48.320 --> 02:36:52.320
Use different branches of technology in a
real

02:36:52.320 --> 02:36:56.320
project and apply the knowledge from the
third

02:36:56.320 --> 02:37:00.000
and fourth years of ESO in a real
project.

02:37:00.000 --> 02:37:01.560
Also

02:37:01.560 --> 02:37:04.440
Currently the. The.

02:37:04.440 --> 02:37:08.440
Robots and humanoids are developing a
lot, and even

02:37:08.440 --> 02:37:11.440
more so with artificial intelligence.

02:37:11.960 --> 02:37:15.960
Examples of this include Tesla robots,
Hanson

02:37:15.960 --> 02:37:19.040
Robotics robots, and figure-e.

02:37:21.360 --> 02:37:25.360
On the other hand, bullying also
continues and is a serious

02:37:25.360 --> 02:37:29.400
problem for educational institutions
around the world.

02:37:30.440 --> 02:37:34.440
According to UNESCO, a 2024 report

02:37:34.440 --> 02:37:38.440
showed that roughly one in three

02:37:38.440 --> 02:37:42.000
students have suffered bullying,

02:37:42.000 --> 02:37:45.000
and this is quite worrying.

02:37:46.240 --> 02:37:49.240
It's not much anymore.

02:37:49.800 --> 02:37:53.120
for the creation of the projection plate.

02:37:53.160 --> 02:37:55.440
We have gone through several processes
and in this video we are going to show

02:37:55.440 --> 02:37:56.600
some of them.

02:37:56.640 --> 02:37:57.800
Well, all of them.

02:37:57.800 --> 02:38:01.800
In the first one we used a translucent,
matte,

02:38:01.800 --> 02:38:05.800
flesh-colored polypropylene plate that we
tried

02:38:05.800 --> 02:38:08.920
to melt with a heat gun on a ceramic
mold.

02:38:09.240 --> 02:38:10.040
This is here.

02:38:10.040 --> 02:38:14.040
However, this did not yield results, as
the

02:38:14.040 --> 02:38:19.240
heat was not distributed evenly and it
did not remove well.

02:38:21.000 --> 02:38:24.880
The second step we went through was very
similar

02:38:24.880 --> 02:38:27.880
with another piece of polypropylene.

02:38:27.920 --> 02:38:31.920
This one is white, translucent, and
matte, which we wanted

02:38:31.920 --> 02:38:36.880
to heat in a home oven so that we could
achieve the same result in an oven.

02:38:36.920 --> 02:38:40.920
The heat was distributed evenly this
time, but we

02:38:40.920 --> 02:38:44.920
encountered some problems, since the
plastic,

02:38:44.920 --> 02:38:50.160
due to its local composition, completely
lost its translucency.

02:38:51.160 --> 02:38:55.080
The matte finish wears off and also
appears in bubbles.

02:38:56.000 --> 02:39:00.000
The third and final step consisted of a
flexible copper

02:39:00.000 --> 02:39:03.000
mesh that we placed over the ceramic mold

02:39:03.000 --> 02:39:06.000
and then applying epoxy resin on top.

02:39:06.000 --> 02:39:10.680
If we can make it harden and adapt to
nobody's shape.

02:39:10.800 --> 02:39:16.600
However, it was no longer translucent and
shone too brightly.

02:39:16.920 --> 02:39:20.920
It's a bit far away for us, but hey,
we're using

02:39:20.920 --> 02:39:24.120
that side for skeletal polypropylene.

02:39:24.120 --> 02:39:27.120
Are you seeing if Enrique can get out?

02:39:27.120 --> 02:39:29.840
No no.

02:39:29.840 --> 02:39:33.840
As a final result, we decided to opt for
this flat

02:39:33.840 --> 02:39:37.840
polypropylene plate with a Verge plastic
edge made

02:39:37.840 --> 02:39:41.560
from recycled bottles, designed in the
Blender

02:39:41.560 --> 02:39:44.560
program and printed on our 3D printer.

02:39:45.320 --> 02:39:49.320
In the end, we decided, using the methods
available to us, to leave

02:39:49.320 --> 02:39:53.520
the face flat using a pink polypropylene
plate framed on a 3D base.

02:39:53.920 --> 02:39:55.160
of PC.

02:39:55.160 --> 02:39:58.960
This way we have guarantees that the
image will be sharp and clear, since

02:39:58.960 --> 02:40:01.960
polypropylene is a material that is quite
translucent.

02:40:13.760 --> 02:40:16.760
Let's see, this is the way to.

02:40:57.960 --> 02:40:58.680
Average.

02:40:58.680 --> 02:41:01.000
Well then, above all, we can take it with
us.

02:41:01.000 --> 02:41:04.000
This plate, of course, this one needs to
be washed more often.

02:41:06.800 --> 02:41:11.880
To allow an optimal image of the humanoid
on the mask.

02:41:11.880 --> 02:41:15.880
Our project had to be at a distance of
between 0.8 meters

02:41:15.880 --> 02:41:19.520
and 2.5 meters, which was an abysmal
distance, so

02:41:19.520 --> 02:41:22.520
we decided to shorten it using lenses.

02:41:23.280 --> 02:41:27.280
We started using diverging lenses, but
these

02:41:27.280 --> 02:41:31.080
didn't work because they stopped the
light and

02:41:31.080 --> 02:41:34.080
left us with a much more distant image.

02:41:34.120 --> 02:41:38.120
So we decided to use converging lenses,
which allow

02:41:38.120 --> 02:41:43.320
us to enlarge the image and therefore
bring it closer to the mask.

02:41:44.080 --> 02:41:48.080
But we still had a very large distance
left,

02:41:48.080 --> 02:41:53.760
so we decided to use a flat mirror that
allowed for projection.

02:41:54.040 --> 02:41:56.320
I'm rambling. Never mind.

02:41:56.320 --> 02:41:58.560
We have the head, cut off the neck.

02:41:58.560 --> 02:42:02.560
How's it going? So, planning

02:42:02.560 --> 02:42:06.360
well, but not just because.

02:42:08.840 --> 02:42:12.840
What you have to do is form a right angle
with

02:42:12.840 --> 02:42:16.960
your eyes closed, using some movements.

02:42:17.080 --> 02:42:18.160
It had to be the same face.

02:42:18.160 --> 02:42:20.480
With the eyes in the same position,
because otherwise it would look like the

02:42:20.480 --> 02:42:22.800
eye was closing lower. That would be very
strange.

02:42:22.800 --> 02:42:25.440
And for example, if we go to everything
related to the mouth, there it requires a

02:42:25.440 --> 02:42:26.800
much

02:42:26.800 --> 02:42:30.800
greater complexity, because it votes for
the same thing, teeth, etc., etc. And we

02:42:30.800 --> 02:42:34.760
also had to have costumes for smiling,
closed mouth and open mouth, naturally.

02:42:34.760 --> 02:42:38.760
So, of course, for the conceptualized AI,
as it

02:42:38.760 --> 02:42:42.760
is a naturally open mouth, we have to
give it a

02:42:42.760 --> 02:42:46.760
rundown through Google Mouse,
cyberbullying,

02:42:46.760 --> 02:42:50.760
cyberbullying, differences, types,
because

02:42:50.760 --> 02:42:54.400
it is responsible for passing the series.

02:42:54.440 --> 02:42:58.440
Change your disguise from a serious mouth
to a smiling mouth to make

02:42:58.440 --> 02:43:02.320
it more dynamic, and then try to think a
little bit and be flexible.

02:43:02.320 --> 02:43:04.520
And we've also added Vale.

02:43:04.520 --> 02:43:08.520
And in the end, well, I'm finishing up,
sorry, it comes out

02:43:08.520 --> 02:43:12.520
there a little bit, it starts to work
when a harassment situation

02:43:12.520 --> 02:43:17.200
appears, as they are in the trial phase
they still have a month left.

02:43:17.200 --> 02:43:21.200
What is more common are insults, mockery,

02:43:21.200 --> 02:43:25.200
pushing or social exclusion and movement,

02:43:25.200 --> 02:43:28.640
which they also achieve.

02:43:31.320 --> 02:43:32.240
It's in the trial period.

02:43:32.240 --> 02:43:35.240
It's a moment that if we leave something
wrong.

02:43:37.040 --> 02:43:39.200
So, if you want to see the whole thing...

02:43:39.200 --> 02:43:42.200
He decided to come to the Tec challenge
on June 16th.

02:43:47.600 --> 02:43:49.320
Close.

02:43:49.320 --> 02:43:50.560
Sorry.

02:43:50.560 --> 02:43:54.560
And here we have Antonio Prieto Sánchez
making a

02:43:54.560 --> 02:43:58.560
fuss, and you see a project that is not
project-based

02:43:58.560 --> 02:44:02.560
learning for the first year of ESO,
Spanish

02:44:02.560 --> 02:44:07.480
Language and Literature, so it's going to
be interesting.

02:44:07.480 --> 02:44:07.960
Thank you so much.

02:44:09.360 --> 02:44:10.520
Well, good afternoon everyone.

02:44:10.520 --> 02:44:13.640
First of all, thank you to the entire
course organization for inviting me to

02:44:13.640 --> 02:44:15.240
give this presentation.

02:44:15.320 --> 02:44:18.000
Secondly, taking into account the time it
is.

02:44:18.000 --> 02:44:19.720
I will try to be as brief as possible.

02:44:19.720 --> 02:44:23.720
Furthermore, all the data is in the
course database, and somewhat

02:44:23.720 --> 02:44:27.440
using the analogy that Gregorio made at
the beginning.

02:44:27.480 --> 02:44:31.480
Well, I think that after having seen
everything we've seen

02:44:31.480 --> 02:44:35.480
here about plates, humanoids, robotics,
etc., etc., I

02:44:35.480 --> 02:44:39.480
think this dish is going to be much
lighter and I also hope to

02:44:39.480 --> 02:44:43.360
get a few laughs, at least within the
auditorium, right?

02:44:43.760 --> 02:44:47.760
The most important thing about this
project is that it is a learning

02:44:47.760 --> 02:44:51.760
situation, because it places the
first-year ESO students of the

02:44:51.760 --> 02:44:55.760
school where I teach the Language and
Literature subject in a real

02:44:55.760 --> 02:44:59.760
context such as Carnival and links it to
the lyric genre and to the

02:44:59.760 --> 02:45:04.560
prior knowledge that we have been giving
throughout the second evaluation.

02:45:04.600 --> 02:45:08.600
But precisely to foster a critical spirit
regarding the advantages

02:45:08.600 --> 02:45:12.600
and disadvantages of artificial
intelligence, that is basically

02:45:12.600 --> 02:45:15.720
the foundation of this learning
situation.

02:45:15.880 --> 02:45:20.160
This is the index that I will follow in
my presentation of the teaching proposal.

02:45:20.840 --> 02:45:24.840
At times I will go much faster, but I
repeat because

02:45:24.840 --> 02:45:28.840
all the detailed information is in the
activity

02:45:28.840 --> 02:45:32.840
sheet and all the evidence, photographs,
images,

02:45:32.840 --> 02:45:36.000
etc. are linked precisely in that sheet.

02:45:36.160 --> 02:45:40.160
Regarding the context, it is important to
note that the

02:45:40.160 --> 02:45:43.840
class, or rather the center where I
teach, is art, and Gasset is a public

02:45:43.840 --> 02:45:45.680
center.

02:45:45.680 --> 02:45:50.040
Specifically, I teach this subject at the
M87 campus.

02:45:50.200 --> 02:45:54.200
It is a venue that specializes in
high-performance

02:45:54.200 --> 02:45:58.200
athletes, specifically and especially in
aquatic

02:45:58.200 --> 02:46:01.800
disciplines, sports, especially water
sports.

02:46:01.840 --> 02:46:05.840
There are two other locations, one is the
headquarters at the Higher

02:46:05.840 --> 02:46:09.440
Sports Council and the main one is at the
Santa Fe headquarters in Puente de los

02:46:09.440 --> 02:46:11.280
Franceses.

02:46:11.400 --> 02:46:14.680
And precisely for this reason it has some
rather particular characteristics that I

02:46:14.680 --> 02:46:16.320
will describe.

02:46:16.400 --> 02:46:20.080
In principle, well, the proposal was
designed for about

02:46:20.080 --> 02:46:23.080
ten or 11 hours, approximately ten, 11
sessions.

02:46:23.080 --> 02:46:27.080
Finally there were 12 and we have them

02:46:27.080 --> 02:46:31.080
in this classroom specifically, because

02:46:31.080 --> 02:46:36.480
here the ratios are in a completely
opposite way.

02:46:36.560 --> 02:46:39.640
Here the funnel widens as the courses

02:46:39.640 --> 02:46:42.640
progress, and we only have 15 students.

02:46:43.000 --> 02:46:48.920
Basically, within the programming, we are
in the second quarter.

02:46:48.920 --> 02:46:52.920
The subject of Spanish Language and
Literature, which

02:46:52.920 --> 02:46:56.920
connects everything that is the lyric
genre and all its

02:46:56.920 --> 02:47:00.920
elements rhyme, types of verses, stanzas,
rhetorical

02:47:00.920 --> 02:47:04.920
resources, etc. And the project is
designed above all

02:47:04.920 --> 02:47:08.600
thinking about the development of the
students' competence.

02:47:08.640 --> 02:47:11.760
The educational objectives are all of
these.

02:47:11.800 --> 02:47:15.800
Basically, we want to put into practice
all the knowledge

02:47:15.800 --> 02:47:20.440
of the lyric genre, but linking it to the
Carnival festival.

02:47:20.880 --> 02:47:24.880
We also value all technological advances,
especially those

02:47:24.880 --> 02:47:28.880
related to the development of artificial
intelligence, while

02:47:28.880 --> 02:47:32.840
at the same time practicing and learning
about generative AI

02:47:32.840 --> 02:47:35.840
tools, both for creating texts, images
and sound.

02:47:36.360 --> 02:47:40.360
Of course, we analyze the biases and
impacts in the process, as we have

02:47:40.360 --> 02:47:44.360
learned in the course, and also in the
end, what we want, above all and

02:47:44.360 --> 02:47:47.680
primarily, is to develop a critical
spirit regarding the pros and cons of

02:47:47.680 --> 02:47:49.360
this technology.

02:47:49.760 --> 02:47:53.520
Why is it a competency-based project?

02:47:53.560 --> 02:47:57.560
Well, linguistic competence, especially
if we had to break it down or

02:47:57.560 --> 02:48:01.240
divide it, is divided into four basic
skills as you know: speaking, listening,

02:48:01.240 --> 02:48:03.080
reading and writing.

02:48:03.120 --> 02:48:07.120
In this case, especially placing special
emphasis on critical evaluation,

02:48:07.120 --> 02:48:11.280
since the various sources from which the
students obtain information.

02:48:11.320 --> 02:48:14.680
And in this specific case, in those that
are

02:48:14.680 --> 02:48:17.680
accessed through artificial intelligence.

02:48:17.680 --> 02:48:21.680
Digital competence, above all, is based
on the advantages

02:48:21.680 --> 02:48:25.680
and disadvantages we find in the digital
age, but also

02:48:25.680 --> 02:48:29.680
here, in the creative use of these tools,
as well as in

02:48:29.680 --> 02:48:33.680
fostering a critical spirit towards them,
and the social

02:48:33.680 --> 02:48:37.280
or personal competence to learn and
learn.

02:48:37.440 --> 02:48:41.440
A little, as you will see in the workshop
mechanics in which the

02:48:41.440 --> 02:48:45.440
project has been developed, in the
autonomy itself, which is

02:48:45.440 --> 02:48:49.160
gradually fostering civic competence in
the students.

02:48:49.200 --> 02:48:49.800
Because?

02:48:49.800 --> 02:48:53.520
Basically because we want to train, or I
think, as a teacher.

02:48:53.520 --> 02:48:55.200
This is a personal premise.

02:48:55.200 --> 02:48:59.200
I always tell them that we have to train
good people, good

02:48:59.200 --> 02:49:03.200
citizens and good future professionals,
and in this case,

02:49:03.200 --> 02:49:07.200
this study of the advantages and
disadvantages of AI

02:49:07.200 --> 02:49:11.200
connects with the development of the
goals of the 2030

02:49:11.200 --> 02:49:15.200
Agenda and in terms of awareness and
cultural expressions,

02:49:15.200 --> 02:49:19.200
because, well, knowledge of the artistic
and cultural

02:49:19.200 --> 02:49:23.200
diversity of Spain, specifically of a
festival as popular

02:49:23.200 --> 02:49:26.680
as Carnival, has a special importance
here in its

02:49:26.680 --> 02:49:29.680
connection with the lyrical genre itself.

02:49:29.880 --> 02:49:32.160
As we mentioned before.

02:49:32.160 --> 02:49:36.160
Regarding the contents, they are all
broken down in the sheet, but

02:49:36.160 --> 02:49:40.880
basically it takes it from these three
content blocks of the subject curriculum.

02:49:40.920 --> 02:49:44.920
Regarding methodology, it is an important

02:49:44.920 --> 02:49:48.920
aspect because several methodologies have

02:49:48.920 --> 02:49:52.920
been used; learning through debate and
Visual

02:49:52.920 --> 02:49:56.920
Thinking have also been very important

02:49:56.920 --> 02:50:00.920
throughout the implementation of the
project,

02:50:00.920 --> 02:50:04.920
but also the scientific method itself,
which

02:50:04.920 --> 02:50:10.120
is ultimately to observe, analyze, and
formulate hypotheses.

02:50:10.160 --> 02:50:14.160
These hypotheses become premises, and
these premises then

02:50:14.160 --> 02:50:18.160
lead to certain diagnostic approaches
that have been implemented

02:50:18.160 --> 02:50:22.160
throughout the entire collaborative
project, because it is

02:50:22.160 --> 02:50:26.160
not a collaborative project, and of
course it has resulted in

02:50:26.160 --> 02:50:30.160
a final product in terms of addressing
diversity, because

02:50:30.160 --> 02:50:34.160
here we have to keep in mind that there
is a common denominator

02:50:34.160 --> 02:50:38.880
and that is that all these students are
high-performance athletes.

02:50:38.880 --> 02:50:42.880
This means that many of them get up at
6:00, arrive at class at 10:30,

02:50:42.880 --> 02:50:47.080
eat in the center's dining room, but then
many go back to training again.

02:50:47.080 --> 02:50:50.080
In other words, they have hardly any time
and their

02:50:50.080 --> 02:50:53.080
weekends are mainly focused on training
or competing.

02:50:53.080 --> 02:50:57.680
Therefore, most of them barely have any
time.

02:50:57.760 --> 02:51:01.760
Added to this is the exhaustion and
fatigue and the great pressures

02:51:01.760 --> 02:51:05.760
they have, not only those of themselves,
their own self-demand, but

02:51:05.760 --> 02:51:09.760
also those that in many cases are exerted
on them by their own coaches,

02:51:09.760 --> 02:51:13.760
clubs or federations, not always with a
very humane treatment,

02:51:13.760 --> 02:51:17.880
something that is not often said and also
those of their own families.

02:51:17.920 --> 02:51:21.920
And this is a very important element to
consider because

02:51:21.920 --> 02:51:25.920
these are students who have real
difficulties in being able

02:51:25.920 --> 02:51:29.680
to learn throughout their normal
educational stage.

02:51:29.680 --> 02:51:32.520
It's very complicated, of course.

02:51:32.520 --> 02:51:36.520
Then there are also some students with
special needs,

02:51:36.520 --> 02:51:42.360
for whom, especially in this case, Visual
Thinking was quite important.

02:51:42.840 --> 02:51:46.840
Regarding the evaluation, different tools
have been

02:51:46.840 --> 02:51:50.840
used, since what is the process we have
followed throughout

02:51:50.840 --> 02:51:54.840
all the sessions, in which there has been
continuous

02:51:54.840 --> 02:51:58.840
interaction with all the students, both
in the lectures,

02:51:58.840 --> 02:52:04.000
but those lectures have also had a lot of
feedback from the students.

02:52:04.000 --> 02:52:08.000
Also, the absorption of the entire
process, the use of digital

02:52:08.000 --> 02:52:12.000
tools, but also the textbook notebook,
the notes and of course,

02:52:12.000 --> 02:52:16.000
the feedback that we have been giving not
only me, but also a

02:52:16.000 --> 02:52:20.000
student who was with me in the practice
of the Master of Teaching

02:52:20.000 --> 02:52:23.080
to guide precisely the work and the
moments.

02:52:23.080 --> 02:52:23.960
Well, everyone.

02:52:23.960 --> 02:52:27.960
The entire project has been scrutinized,
every single moment examined closely; the

02:52:27.960 --> 02:52:30.760
whole process has been evaluated through
everything they have been doing little by

02:52:30.760 --> 02:52:32.200
little.

02:52:32.200 --> 02:52:36.160
It has been very structured and organized
in workshops and we have obviously also

02:52:36.160 --> 02:52:38.160
valued its evolution.

02:52:38.640 --> 02:52:41.640
Elements that we have all valued are in
this rubric.

02:52:41.960 --> 02:52:43.800
Basically, these have been the four
elements, although others were introduced

02:52:43.800 --> 02:52:44.760
later.

02:52:44.760 --> 02:52:47.400
We introduced a change that I will
explain later, which is teamwork, the

02:52:47.400 --> 02:52:48.760
process

02:52:48.760 --> 02:52:52.760
and organization, the adaptation to the
content, that is, that they have

02:52:52.760 --> 02:52:56.760
followed everything that has been passed
on to them and the oral expression

02:52:56.760 --> 02:53:00.480
which is the final presentation of their
projects, both orally and in a

02:53:00.480 --> 02:53:03.480
presentation that they had to design in
the last session.

02:53:03.600 --> 02:53:07.600
Regarding the schedule, from now on I
present the presentation

02:53:07.600 --> 02:53:11.200
I made, which is the backbone of the
entire project throughout all the

02:53:11.200 --> 02:53:13.000
sessions.

02:53:13.000 --> 02:53:15.640
This is what they have been seeing; she
tried to summarize it at times, but you

02:53:15.640 --> 02:53:17.000
must

02:53:17.000 --> 02:53:20.280
take into account, especially regarding
the circumstances of these students and

02:53:20.280 --> 02:53:21.960
also of the center.

02:53:21.960 --> 02:53:22.520
Very important.

02:53:22.520 --> 02:53:26.520
It's a public school, but we don't have a
gym, we hardly have any

02:53:26.520 --> 02:53:30.520
recess, we don't have a computer room, we
don't have a library, the

02:53:30.520 --> 02:53:34.520
teachers' room is tiny, and you have to
use computers in this school

02:53:34.520 --> 02:53:37.640
because there's no technology room by
using a trolley.

02:53:37.640 --> 02:53:41.200
When they are not disconnected, the
operating system does not work or they

02:53:41.200 --> 02:53:44.200
connect to another operating system that
they do not know.

02:53:44.200 --> 02:53:48.200
Or basically put it in the car, take the
car down because it's on

02:53:48.200 --> 02:53:52.200
the second floor, and go down to ours
which is on the first floor.

02:53:52.200 --> 02:53:56.200
It's also a quixotic task. Often, putting
a computer in

02:53:56.200 --> 02:54:00.120
the classroom is a completely risky
undertaking because

02:54:00.120 --> 02:54:03.120
you never know what's going to happen.

02:54:03.160 --> 02:54:07.160
Therefore, I want to say that we have to
forget about technology rooms here, and

02:54:07.160 --> 02:54:10.360
obviously, after everything that has been
seen here, it is a completely different

02:54:10.360 --> 02:54:12.000
context.

02:54:12.480 --> 02:54:15.480
What do we do in the first session?

02:54:15.800 --> 02:54:17.160
There are basically six workshops.

02:54:17.160 --> 02:54:20.480
The first is the presentation, in which
we try above all

02:54:20.480 --> 02:54:23.480
to introduce the students to the world of
Carnival.

02:54:23.520 --> 02:54:24.720
What is Carnival?

02:54:24.720 --> 02:54:28.720
And in the end, this is the underlying
principle and umbrella concept

02:54:28.720 --> 02:54:32.600
of the project: that Carnival and poetry
are exactly the same thing.

02:54:33.240 --> 02:54:37.240
I am gradually introducing the students
to the slang of carnival,

02:54:37.240 --> 02:54:40.760
to the different terms they need to be
aware of in order to optimally develop

02:54:40.760 --> 02:54:42.560
this project.

02:54:42.680 --> 02:54:45.680
We are, therefore, obviously talking
about the history of Carnival.

02:54:45.680 --> 02:54:46.840
Where does it come from?

02:54:46.840 --> 02:54:51.680
What is a drop, what is a murga, what is
a comparsa?

02:54:51.680 --> 02:54:55.480
Why has Carnival become such a popular
and

02:54:55.480 --> 02:54:58.480
important festival in those anthems?

02:54:58.520 --> 02:54:59.960
Well, it's not available here.

02:54:59.960 --> 02:55:03.960
It was basically Celia Cruz's carnival,
the song "Y ya

02:55:03.960 --> 02:55:08.160
vez que intromos estas conceptos vamos a
dar la murga".

02:55:08.200 --> 02:55:08.800
Because?

02:55:08.800 --> 02:55:12.800
Because what we really want is for our
voice to be heard, and we're

02:55:12.800 --> 02:55:15.840
going to try, even though we're talking
about

02:55:15.840 --> 02:55:18.840
disadvantages or advantages, to have a
good time.

02:55:18.840 --> 02:55:21.840
What do all these carnival groups have in
common?

02:55:21.880 --> 02:55:25.920
Well, they're aiming for a parody, a
social satire of some current event.

02:55:26.200 --> 02:55:28.720
And what is it?

02:55:28.720 --> 02:55:32.720
What exactly are these groups made up of,
which in

02:55:32.720 --> 02:55:36.720
different parts of Spain hold
competitions that are

02:55:36.720 --> 02:55:40.720
more or less half-hour performances in
which they

02:55:40.720 --> 02:55:44.720
dedicate a lot of time and money and
which are also

02:55:44.720 --> 02:55:48.720
festivals of great tourist and cultural
interest?

02:55:48.720 --> 02:55:52.000
Well, we explain that all these parts are
the ones in which these voices are

02:55:52.000 --> 02:55:53.680
divided.

02:55:53.760 --> 02:55:54.760
What are they going to do?

02:55:54.760 --> 02:55:58.760
They're going to do a medley of choruses,
basically selecting three

02:55:58.760 --> 02:56:01.880
choruses from songs and adapting them to
a chosen theme.

02:56:02.080 --> 02:56:03.600
But we give many examples.

02:56:03.600 --> 02:56:04.760
Here's the visual.

02:56:04.760 --> 02:56:08.600
Look, I only want to reproduce this one
because it seems representative to me,

02:56:08.600 --> 02:56:11.600
especially because of its relation to the
subject, you know.

02:56:19.880 --> 02:56:25.080
I live on a planet that speaks our
grandmother's language.

02:56:26.800 --> 02:56:31.720
Languages ​​of Jaén, different languages
​​with their own identity.

02:56:33.760 --> 02:56:37.760
They told me at school that these are

02:56:37.760 --> 02:56:41.760
super important off-hours and that

02:56:41.760 --> 02:56:45.760
nobody has the right to impose their

02:56:45.760 --> 02:56:51.280
way of speaking, erasing the message of
freedom.

02:56:51.280 --> 02:56:54.280
Ladders.

02:56:54.960 --> 02:56:58.960
Staircase in Cancun, in his shop in
Galegos

02:56:58.960 --> 02:57:02.960
to a seafaring countryside and the words

02:57:02.960 --> 02:57:06.960
Sing la, la, la, la la freedom and they
will

02:57:06.960 --> 02:57:10.120
die and the language of many

02:57:10.120 --> 02:57:13.120
generations of my planet.

02:57:13.120 --> 02:57:17.120
People get lost,

02:57:17.120 --> 02:57:21.800
the hearts of many.

02:57:24.800 --> 02:57:28.800
This is an example, for instance, of that
awareness and cultural expressions, not

02:57:28.800 --> 02:57:32.800
only regarding the linguistic diversity
of Spain, but also because they have to

02:57:32.800 --> 02:57:36.040
promote those values ​​of respect for our
linguistic diversity.

02:57:36.080 --> 02:57:40.080
But at the same time, as you can see, the
introduction also gradually

02:57:40.080 --> 02:57:44.080
introduces them to the concept of type
and the important thing, which is

02:57:44.080 --> 02:57:48.080
not just the lyrics, because in fact we
tell them that there are many people

02:57:48.080 --> 02:57:52.080
in the López de Ayala theater or in the
COAC of the Falla Theater of Cádiz,

02:57:52.080 --> 02:57:55.600
who dedicate themselves all year simply
to living precisely from these lyrics

02:57:55.600 --> 02:57:57.400
that they compose.

02:57:57.400 --> 02:58:01.400
For these groups, and throughout many
examples we

02:58:01.400 --> 02:58:05.400
give, there are different types, for
example, in the

02:58:05.400 --> 02:58:09.400
introduction of this chirigota we simply
see how

02:58:09.400 --> 02:58:14.320
important the staging is and of course
also the non-verbal language.

02:58:14.800 --> 02:58:17.040
I'm only including the beginning of it
because we can't.

02:58:22.600 --> 02:58:26.600
I live with a couple who are very

02:58:26.600 --> 02:58:30.320
naughty in bed and often shout

02:58:30.320 --> 02:58:33.320
"Antonio" at many nights.

02:58:34.320 --> 02:58:38.320
David is making a child, Teresa, please
don't

02:58:38.320 --> 02:58:43.280
make any more children, they won't all
fit on the table.

02:58:45.800 --> 02:58:49.800
Mrs. Bueno, there are many examples
collected from

02:58:49.800 --> 02:58:53.440
different Carnival celebrations
throughout Spain.

02:58:53.480 --> 02:58:56.400
More recent ones are fine, less recent
ones too.

02:58:56.400 --> 02:59:00.400
We also play them some versions of
well-known songs about this

02:59:00.400 --> 02:59:04.400
hostel phenomenon, although they're
talking about it all day

02:59:04.400 --> 02:59:08.400
long, and then, once we've covered the
specifics of Carnival

02:59:08.400 --> 02:59:12.320
and these groups in this presentation,
we'll explain what they need to do at

02:59:12.320 --> 02:59:14.280
this point.

02:59:14.280 --> 02:59:17.280
Put yourselves in context, they only have
one.

02:59:17.440 --> 02:59:20.080
Basically, their faces are a picture, and
it's a picture because they're both

02:59:20.080 --> 02:59:21.440
waiting to

02:59:21.440 --> 02:59:24.200
find out if they're going to have to make
a stop in this direction because they

02:59:24.200 --> 02:59:25.600
don't have time.

02:59:25.600 --> 02:59:27.480
And two, knowing with whom.

02:59:27.480 --> 02:59:29.720
Since they know it's a team project, they
need to know who they'll have to meet

02:59:29.720 --> 02:59:30.880
with.

02:59:30.920 --> 02:59:34.800
And that, of course, is basically his
only concern at the moment.

02:59:34.800 --> 02:59:38.800
So, well, it is explained to them what a
project is, that

02:59:38.800 --> 02:59:42.800
is, a collaborative project based on a
final product

02:59:42.800 --> 02:59:46.800
in which they have to relate and put into
practice all

02:59:46.800 --> 02:59:50.800
the knowledge about the lyrical genre
that they will

02:59:50.800 --> 02:59:54.800
work on in teams and that in the end the
objective is that

02:59:54.800 --> 02:59:58.800
they compose or adapt the songs of three
choruses of

02:59:58.800 --> 03:00:02.800
popular songs to a theme that they choose
through a

03:00:02.800 --> 03:00:06.800
workshop that we will be doing at the
beginning, which

03:00:06.800 --> 03:00:09.920
will help them decide on what the theme
is

03:00:09.920 --> 03:00:12.920
and of course they are shown the rubric.

03:00:13.040 --> 03:00:17.040
And obviously the important thing here is
basically that all they

03:00:17.040 --> 03:00:21.040
have to do is choose the three, adapt and
rewrite the lyrics, but

03:00:21.040 --> 03:00:25.040
they have to relate it to a theme of how
the development of artificial

03:00:25.040 --> 03:00:29.040
intelligence affects the Sustainable
Development Goals of 2030,

03:00:29.040 --> 03:00:33.040
which they will then set to music with
their own song, and then, in

03:00:33.040 --> 03:00:37.040
addition to that lyric in which they also
have to incorporate and

03:00:37.040 --> 03:00:41.960
respect the meter and rhyme approximately
of those choruses, they have to.

03:00:44.920 --> 03:00:47.560
They have to edit that lyric
audiovisually using another tool, like a

03:00:47.560 --> 03:00:48.920
video,

03:00:48.920 --> 03:00:52.920
and then they have to present it, they
all have to present and explain it,

03:00:52.920 --> 03:00:56.360
participating more or less at the same
time and taking into account all the

03:00:56.360 --> 03:00:58.120
requirements they have to meet.

03:00:58.200 --> 03:01:02.200
They show them, for example, a video from
one of

03:01:02.200 --> 03:01:06.200
the Muga del Temple series, in this case
the one

03:01:06.200 --> 03:01:10.200
from Gran Canaria, in which, well, if I
go to the

03:01:10.200 --> 03:01:14.200
chorus, what I do with all the waste, the
Luis

03:01:14.200 --> 03:01:18.200
Fonsi song, everything in its place, I
put, you

03:01:18.200 --> 03:01:22.200
said, if cans in the yellow bin and the
green

03:01:22.200 --> 03:01:26.200
container only for glass, I put all the
little

03:01:26.200 --> 03:01:30.200
papers in the blue bin and if I see that
you don't

03:01:30.200 --> 03:01:34.200
do the same I'm going to force you to
pick up little

03:01:34.200 --> 03:01:37.840
pieces. I'm passing by so that you don't.

03:01:37.840 --> 03:01:40.120
Okay, we formed the teams, we
communicated it.

03:01:40.120 --> 03:01:43.280
Here, angry faces relax, some even show
joy, and it's explained to everyone in

03:01:43.280 --> 03:01:44.880
the workshops.

03:01:44.880 --> 03:01:48.080
So far we have been at the project
presentation.

03:01:48.120 --> 03:01:51.160
We now continue with the research
workshop.

03:01:51.200 --> 03:01:55.200
Then we have the Creative Writing
workshop, followed by music

03:01:55.200 --> 03:01:59.200
production, the audiovisual editing
workshop, and finally

03:01:59.200 --> 03:02:02.560
the performances related to the research
workshop.

03:02:02.560 --> 03:02:06.040
We'll start, and you'll all recognize
this quickly, by talking about what AI

03:02:06.040 --> 03:02:07.800
is.

03:02:07.840 --> 03:02:11.440
Well, I'm following the process we saw in
the course in the first place.

03:02:11.440 --> 03:02:15.440
Well, first, I'll refer to its definition
in the RAE (Royal

03:02:15.440 --> 03:02:20.440
Spanish Academy) and clarify that we are
talking about a computer language.

03:02:20.760 --> 03:02:24.760
Later I talk to them about the main
technological resources and technological

03:02:24.760 --> 03:02:28.600
advances that have given rise to those
foundational models like GPT, which are

03:02:28.600 --> 03:02:30.560
the ones they know.

03:02:30.840 --> 03:02:34.840
And we talk about those concepts like
neural networks and algorithms

03:02:34.840 --> 03:02:38.480
at the same time, because I also
illustrate it with very simple things

03:02:38.480 --> 03:02:40.320
that they understand.

03:02:40.320 --> 03:02:44.320
I don't know if they look very good, but
well, in the morphological

03:02:44.320 --> 03:02:48.320
analysis guidelines we give in language,
I do a bit of that neural

03:02:48.320 --> 03:02:52.320
network in which we are faced with a
variable word, yes or no, and that

03:02:52.320 --> 03:02:56.000
variable word has a degree, if so, an
adjective no, well, the rest.

03:02:56.000 --> 03:02:59.000
I mean, so that you fully understand how
it works.

03:02:59.000 --> 03:03:03.000
But also, well, we apply machine
learning, that is,

03:03:03.000 --> 03:03:08.440
we do, you know, the task we've seen
about cats, dogs, white, black.

03:03:08.840 --> 03:03:12.840
We talked about the basic processes of
artificial intelligence: training,

03:03:12.840 --> 03:03:16.600
learning, and testing, which they, as
athletes, understand very well.

03:03:16.640 --> 03:03:18.560
We tried the Aquinate for a moment.

03:03:18.560 --> 03:03:22.400
Later we talked about a good pron, and
how important it is to

03:03:22.400 --> 03:03:25.400
obtain the most reliable information
possible.

03:03:25.400 --> 03:03:29.400
Although we have biases and impacts, we
relate these to the elements of

03:03:29.400 --> 03:03:33.400
communication, which is important and
something you've already seen since

03:03:33.400 --> 03:03:37.800
the beginning of the course. And finally,
we also discuss biases and their impacts.

03:03:37.800 --> 03:03:40.640
So here we simply discuss the topic of
professions.

03:03:40.640 --> 03:03:44.640
As you know, they showed us this in the
course,

03:03:44.640 --> 03:03:48.040
and this really caught his attention.

03:03:48.080 --> 03:03:52.080
The goal is basically to then establish a
debate about these

03:03:52.080 --> 03:03:56.080
2030 Agenda goals and see what the pros
and cons are regarding

03:03:56.080 --> 03:03:59.160
artificial intelligence in relation to
them.

03:03:59.200 --> 03:04:02.920
We write down everything they create on
the board, we discuss it, and we resolve

03:04:02.920 --> 03:04:04.800
any doubts at the end.

03:04:04.840 --> 03:04:07.480
Well, we've come to this sort of thing.

03:04:07.480 --> 03:04:10.840
We talked, I insist, about both the
advantages and the disadvantages.

03:04:10.880 --> 03:04:13.920
But what happened, and what was
surprising about the project?

03:04:13.960 --> 03:04:17.960
So basically, they finally decided to
talk about the disadvantages in all their

03:04:17.960 --> 03:04:21.960
groups, perhaps because they were the
ones they understood the most, or perhaps

03:04:21.960 --> 03:04:25.960
because in the professional sphere the
advantages when they go months and they,

03:04:25.960 --> 03:04:29.480
well, for them it is an advantage, is
being able to copy through AI, but

03:04:29.480 --> 03:04:32.480
they know that it doesn't bring them
anything good, of course.

03:04:32.480 --> 03:04:36.480
In this first creative writing workshop
they don't have

03:04:36.480 --> 03:04:40.480
any technological tools, we don't have
computers and I

03:04:40.480 --> 03:04:44.480
simply ask them to think about what theme
they want to decide

03:04:44.480 --> 03:04:48.600
on and they have to choose three choruses
from popular songs.

03:04:48.640 --> 03:04:52.640
I also insist that it's not that they
shouldn't choose reggaeton

03:04:52.640 --> 03:04:56.640
or any music at all, because normally
these lyrics, well,

03:04:56.640 --> 03:05:00.120
adapting them isn't so easy and they
write them down.

03:05:00.120 --> 03:05:02.680
It consists of three creative writing
workshop sessions.

03:05:02.680 --> 03:05:06.000
In the first one, the goal is basically
for them to at least manage to adapt a

03:05:06.000 --> 03:05:07.680
chorus.

03:05:08.400 --> 03:05:12.400
Once they have adapted that refrain, we
move forward and

03:05:12.400 --> 03:05:16.400
through the Notebook LM tool with an
account password

03:05:16.400 --> 03:05:20.400
that we created directly with the intern,
we create four

03:05:20.400 --> 03:05:24.400
notebooks, each one has a notebook and in
the second session

03:05:24.400 --> 03:05:28.400
I offer each group a computer with the
pro maestro that you

03:05:28.400 --> 03:05:34.280
will remember that they offered us in the
course, but adapted to this project.

03:05:34.320 --> 03:05:38.320
Well, it's not here, I haven't included
it, but of course, in the one

03:05:38.320 --> 03:05:42.320
where they have to behave like carnival
troupes, basically you have

03:05:42.320 --> 03:05:46.320
to give them the context, what the role
is, etc. All the important

03:05:46.320 --> 03:05:50.320
elements of a pron and to my surprise, in
the second session you start

03:05:50.320 --> 03:05:53.560
to hear how when they interact with the
tool and are critical

03:05:53.560 --> 03:05:56.560
of it as most of the options that
Notebook proposes.

03:05:56.560 --> 03:06:00.880
LM because they do not conform to the
syllabic scheme of the original refrains.

03:06:00.880 --> 03:06:02.920
As for what they want, there are things I
don't understand.

03:06:02.920 --> 03:06:03.760
Start listening.

03:06:03.760 --> 03:06:07.480
To my surprise, the AI ​​doesn't listen
to me. The AI ​​is useless. It's much

03:06:07.480 --> 03:06:09.360
more fun if we do it ourselves.

03:06:09.360 --> 03:06:12.120
Why would we use AI if we can do it
faster ourselves?

03:06:12.120 --> 03:06:16.120
In short, except for one group out of the
four I formed,

03:06:16.120 --> 03:06:20.320
basically everyone adapted the lyrics in
their own handwriting.

03:06:20.600 --> 03:06:24.600
Later, once we had all those letters,
because I controlled

03:06:24.600 --> 03:06:28.600
this creative writing process a lot, then
they had to leave

03:06:28.600 --> 03:06:32.280
everything on a stick, on a pen, they
couldn't take it home.

03:06:32.280 --> 03:06:34.840
He also didn't want them to do anything
at home; everything had to be done in the

03:06:34.840 --> 03:06:36.120
classroom.

03:06:36.200 --> 03:06:40.760
So we started with the tool once we had
all the lyrics to set to music here.

03:06:40.760 --> 03:06:44.760
The milestone was that they would ask for
different versions of

03:06:44.760 --> 03:06:48.760
different musical genres: electronic,
pop, rock, Spanish music,

03:06:48.760 --> 03:06:52.720
flamenco, that is, whatever they wanted,
but they had to do tests with different

03:06:52.720 --> 03:06:54.720
musical genres.

03:06:54.720 --> 03:06:57.600
And then I'll tell you what the surprise
of the elections was.

03:06:57.600 --> 03:06:59.640
And finally, the audiovisual editing
workshop.

03:06:59.640 --> 03:07:05.000
Once they had their choruses set to
music, I used the video tool.

03:07:05.000 --> 03:07:09.000
Today I already knew that I wasn't going
to get any performance,

03:07:09.000 --> 03:07:13.000
obviously, like the one we saw at the
Manuel de Falla or López

03:07:13.000 --> 03:07:17.000
de Ayala theaters, it was impossible, but
I wanted them to

03:07:17.000 --> 03:07:21.000
realize that the artificial intelligence
tool was obviously

03:07:21.000 --> 03:07:25.040
always much more powerful if they had a
paid version.

03:07:25.080 --> 03:07:29.080
In fact, when they used their credits
here in video there or checked

03:07:29.080 --> 03:07:33.000
that they all had the same images in all
the letters, well, they

03:07:33.000 --> 03:07:36.000
said that this is a deception, this is a
scam.

03:07:36.000 --> 03:07:36.720
How can that be?

03:07:36.720 --> 03:07:39.320
But that was the goal I intended with
this.

03:07:39.320 --> 03:07:40.400
And what happened in the end?

03:07:40.400 --> 03:07:44.400
Well, I announced here that I was going
to

03:07:44.400 --> 03:07:48.400
give an extra point in the rubric to
those

03:07:48.400 --> 03:07:53.960
teams that chose that letter which they
then adapted to music.

03:07:55.520 --> 03:07:56.920
Act like a drop.

03:07:56.920 --> 03:08:00.920
In other words, they didn't just have to
talk while

03:08:00.920 --> 03:08:04.920
the music played, or do some kind of
crime or performance

03:08:04.920 --> 03:08:09.160
to put them in context about what he had
decided.

03:08:09.160 --> 03:08:12.200
Because there was one student who didn't
want us to

03:08:12.200 --> 03:08:15.200
offer it, whose mother didn't want us to
offer it.

03:08:15.440 --> 03:08:19.440
I didn't ask the family for permission,
but the

03:08:19.440 --> 03:08:24.480
important thing at the end of the project
was the results.

03:08:24.480 --> 03:08:26.520
And what results have we had?

03:08:26.520 --> 03:08:27.840
Well, from the four teams.

03:08:27.840 --> 03:08:32.040
To give you an idea, I'm sometimes able
to figure out the first team.

03:08:32.400 --> 03:08:35.720
His carnival group or chirigota was
called Los Copian and his theme song

03:08:35.720 --> 03:08:37.400
was...

03:08:37.400 --> 03:08:42.680
The main issue has been the misuse of AI
for copying.

03:08:42.680 --> 03:08:46.280
I'm not going to tell you what the
choruses they used are, but

03:08:46.280 --> 03:08:49.280
I will see if someone divinely revealed
it later.

03:08:50.000 --> 03:08:52.120
This is the result, okay?

03:08:58.440 --> 03:09:02.880
Ti. I copied the answers

03:09:03.240 --> 03:09:06.240
from my snitch,

03:09:06.680 --> 03:09:12.040
I copied and Olivia got caught. I copied.

03:09:12.040 --> 03:09:16.040
He caught me with the paper in

03:09:16.040 --> 03:09:20.040
my hands, I copied it and I only

03:09:20.040 --> 03:09:23.960
got a failing grade because.

03:09:23.960 --> 03:09:27.480
And my mother covered me with sticks.

03:09:29.000 --> 03:09:32.840
I copied and copied, and the conversation
never left me.

03:09:34.600 --> 03:09:40.240
I copied and copied, and I hadn't
reminded her of the argument.

03:09:40.240 --> 03:09:43.240
Another song.

03:09:45.160 --> 03:09:50.080
I'll turn them to stone to pick up my
train.

03:09:50.440 --> 03:09:54.440
And even though the teachers are

03:09:54.440 --> 03:09:59.280
good at this, I'm always the one who
copies.

03:09:59.600 --> 03:10:03.200
But replicating in order to keep

03:10:03.200 --> 03:10:06.200
changing never works out well.

03:10:07.080 --> 03:10:12.720
I will endure the punishments and I will
never give up.

03:10:12.760 --> 03:10:17.120
And even if the exam falls apart, I'll
repeat it.

03:10:18.280 --> 03:10:21.280
I will repeat.

03:10:24.200 --> 03:10:28.440
I'm going to use this template with the
cheat sheet.

03:10:28.680 --> 03:10:31.120
Don't let them catch me off guard by
trusting them.

03:10:31.120 --> 03:10:33.920
I am like the trumpet that deceives me.

03:10:33.920 --> 03:10:36.560
The same note and I want to try it.

03:10:36.560 --> 03:10:40.560
I'm so sad about all this biology, I want
a little

03:10:40.560 --> 03:10:44.560
bit of ease, but look at me and tell me,
teacher,

03:10:44.560 --> 03:10:48.160
I can't see my grade and I can't pass.

03:10:51.720 --> 03:10:56.440
I will remember the argument in front of
everyone.

03:10:57.320 --> 03:11:00.640
I'll turn them to stone.

03:11:01.160 --> 03:11:04.920
Okay, you can listen to it in full within
the virtual

03:11:04.920 --> 03:11:07.920
classroom, but basically here they have
used.

03:11:07.920 --> 03:11:13.440
Se fue de Laura Pausini Resistiré del dúo
anímico y Zapatilla del Canto del Loco.

03:11:13.480 --> 03:11:16.280
This is actually the longer version
because they decided to lengthen it.

03:11:16.280 --> 03:11:18.400
The others are just over a minute long.

03:11:18.400 --> 03:11:22.400
I say this because our humans have
already risen up here. So in the

03:11:22.400 --> 03:11:26.400
next chirigota, "The Soulless Friend,"
the chosen theme was basically

03:11:26.400 --> 03:11:30.400
the dehumanization of AI, and from my
point of view it was one of the

03:11:30.400 --> 03:11:33.560
most interesting, especially because many
understood

03:11:33.560 --> 03:11:36.560
that in the end there was a human behind
them.

03:11:36.560 --> 03:11:41.000
And I am, well, obviously, a Spanish
language and literature expert.

03:11:41.000 --> 03:11:44.600
What I want is for them to go inside and
put aside their

03:11:44.600 --> 03:11:47.600
computers or mobile devices all day long.

03:11:47.600 --> 03:11:51.600
And this is what they did at the end of

03:11:51.600 --> 03:11:56.320
this song: they name the choruses they
used.

03:12:15.800 --> 03:12:19.800
Give your code a little energy,

03:12:19.800 --> 03:12:23.800
because you are not human and you

03:12:23.800 --> 03:12:27.800
don't have a soul, you don't have

03:12:27.800 --> 03:12:30.840
life and you don't feel empathy.

03:12:30.840 --> 03:12:36.640
Cold.

03:12:39.680 --> 03:12:43.680
It has neither consciousness nor a

03:12:43.680 --> 03:12:47.680
little bit of being, nor can it hurt you.

03:12:49.880 --> 03:12:50.400
Tell me.

03:12:50.400 --> 03:12:54.280
And both were, "What do you intend to
do?"

03:12:54.960 --> 03:12:57.960
deceiving.

03:12:59.280 --> 03:13:00.600
You're welcome.

03:13:00.600 --> 03:13:04.160
The con artist is going to use artificial

03:13:04.160 --> 03:13:07.160
intelligence, an international deal.

03:13:07.520 --> 03:13:11.040
A friend who isn't real is in the AI.

03:13:11.040 --> 03:13:14.040
Everyone trusts him.

03:13:14.840 --> 03:13:18.680
It's in AI or we haven't done it.

03:13:18.680 --> 03:13:24.360
Beware of La Macarena by Los del Río and
Yo de la Luna by Mecano.

03:13:24.360 --> 03:13:27.360
Rosalía's pearl.

03:13:33.000 --> 03:13:34.040
I'm putting on the third one.

03:13:34.040 --> 03:13:39.000
We left the 4th one so that it looks good
enough to be discovered in the database.

03:13:39.040 --> 03:13:43.040
In this case, they were called the
professional leagues, and the theme

03:13:43.040 --> 03:13:47.040
they used was basically that the AI ​​is
trained to agree with us. As you

03:13:47.040 --> 03:13:51.040
can see here, since they have done all
the presentations, they show us

03:13:51.040 --> 03:13:55.040
both the adapted and original songs with
their metric scheme, the

03:13:55.040 --> 03:13:58.400
justification of the project, the
problems they encountered, and how they

03:13:58.400 --> 03:14:00.080
solved them.

03:14:00.080 --> 03:14:03.960
And here we have the potpourri

03:14:05.720 --> 03:14:08.720
of this.

03:14:28.240 --> 03:14:32.920
I've been through life and I'm not going
back.

03:14:33.600 --> 03:14:37.880
And seven here, Piti, I'm not going back
in.

03:14:38.640 --> 03:14:43.000
I believed everything like a child and
I'm not getting involved again.

03:14:43.640 --> 03:14:48.120
They all agreed with me, I'm not getting
involved again.

03:14:48.160 --> 03:14:50.680
He tells me that everything is rosy.

03:14:50.680 --> 03:14:53.200
I tell him not to say that to me.

03:14:53.200 --> 03:14:56.120
He tells me he doesn't mind lying to me.

03:14:56.120 --> 03:14:59.240
I'm always right, he won't sway me.

03:15:02.680 --> 03:15:06.360
And if I approach the daughter and she

03:15:06.360 --> 03:15:09.360
lies to me, if she agrees with me.

03:15:10.200 --> 03:15:12.840
What can I do?

03:15:12.840 --> 03:15:15.240
The world is catching me off guard.

03:15:15.240 --> 03:15:19.240
If I create a public square and have

03:15:19.240 --> 03:15:24.760
clouded ideas that are useless, that's
it.

03:15:24.760 --> 03:15:26.880
It's basically just guitar playing, so
it's not important.

03:15:26.880 --> 03:15:30.880
And well, the conclusions of the project,
which are the most

03:15:30.880 --> 03:15:34.880
important for me, are that I have had to
give it more time, something

03:15:34.880 --> 03:15:38.160
I already anticipated, but also because
they have

03:15:38.160 --> 03:15:41.160
been very involved throughout the entire
project.

03:15:41.200 --> 03:15:45.200
I was also surprised that they
themselves, moreover, through all that

03:15:45.200 --> 03:15:49.200
deductive reasoning, after the analysis
in the research of what AI is,

03:15:49.200 --> 03:15:52.600
have decided precisely to talk about the
disadvantages.

03:15:53.000 --> 03:15:58.680
They also noted that the use of prompts
has caught their attention.

03:15:58.680 --> 03:16:01.760
Is it something they didn't do because
many of them have copied me, in fact in

03:16:01.760 --> 03:16:03.320
class, and I've realized why?

03:16:03.320 --> 03:16:07.320
Because the register in which they spoke
was not at all appropriate

03:16:07.320 --> 03:16:11.320
to what they are used to, and also
because the biases in the

03:16:11.320 --> 03:16:15.320
presentation that we have identified in
the presentation of

03:16:15.320 --> 03:16:19.320
all the projects have been clear, because
when putting them all

03:16:19.320 --> 03:16:23.320
together and all the musical registers
that they had asked the

03:16:23.320 --> 03:16:28.440
tool for, basically they have all chosen
Latin rhythms, flamenco or reggaeton.

03:16:28.440 --> 03:16:28.920
Because?

03:16:28.920 --> 03:16:32.760
Because clearly, Castilian music, for
these tools, is surely that.

03:16:32.760 --> 03:16:36.120
It's not because it's not rock or other
echoes like

03:16:36.120 --> 03:16:39.120
not to Castilianize these GPTs, these
GPTs.

03:16:39.240 --> 03:16:43.240
So some of these things have also made
them perfectly

03:16:43.240 --> 03:16:47.240
aware of what these biases and impacts
are, which was

03:16:47.240 --> 03:16:50.520
also one of the objectives of this
project.

03:16:50.520 --> 03:16:53.200
So, well, that's all, and thanks for
listening.

03:16:53.200 --> 03:16:56.200
Thank you.

03:16:57.480 --> 03:17:00.480
I'm not going to put you on.

03:17:00.760 --> 03:17:01.400
That's it, that's it.

03:17:02.600 --> 03:17:03.800
Very good.

03:17:03.800 --> 03:17:06.200
Let's see, let's see, let's see.

03:17:06.200 --> 03:17:09.200
I'll leave it here later, if you want.

03:17:10.480 --> 03:17:13.160
I'm going to close quickly because we're
already outside.

03:17:13.160 --> 03:17:16.360
Out of time, and with this,

03:17:16.360 --> 03:17:19.360
we finish the course.

03:17:19.400 --> 03:17:20.800
I think it's already pending.

03:17:20.800 --> 03:17:24.800
All that remains is for us to give the
signature sheets to our colleagues

03:17:24.800 --> 03:17:28.800
in the Community of Madrid. You don't
need to sign them on the way out,

03:17:28.800 --> 03:17:32.360
to confirm, you don't need to sign the
signature sheet, and

03:17:32.360 --> 03:17:35.360
they will send you the certificate at
some point.

03:17:35.640 --> 03:17:36.640
So I wait.

03:17:36.640 --> 03:17:38.880
I hope you enjoyed this course.

03:17:38.880 --> 03:17:42.440
As you know, we try to make it as diverse
as possible in terms of ages and

03:17:42.440 --> 03:17:44.240
subjects.

03:17:44.400 --> 03:17:48.400
That's complicated, but at least you've

03:17:48.400 --> 03:17:52.400
taken away a few key ideas and that will

03:17:52.400 --> 03:17:56.000
make your classes as fun as Antonio's.

03:17:56.520 --> 03:17:58.080
I wasn't that lucky.

03:17:59.360 --> 03:18:00.080
Literature.

03:18:00.080 --> 03:18:03.320
University was much more traditional, so
that's how it is.