YT YouTube2Text

YouTube2Text → Transcripts → Podcast 9.1 | Big Techs e a influência nas eleições 2026 com o Prof. George Valença | 06/11/2026

Podcast 9.1 | Big Techs e a influência nas eleições 2026 com o Prof. George Valença | 06/11/2026 — Transcript

by TV Guararapes Oficial · 9,450 words · 1,527 segments · language en · Watch on YouTube

Open in transcriber Summarize with AI Download SRT/VTT

Full transcript

0:00Hello,
0:08a very warm welcome to you all to
0:10podcast 9.1, a product here from TV
0:12Guararapes.
0:14We are starting another episode today,
0:18Friday, November 6th, 2026.
0:21The year went by way too fast, the
0:24election is over, thank God, but we are
0:27still going to discuss quite a bit,
0:30including in today's episode, about the
0:332026 elections.
0:35Vandex Santiago, welcome back.
0:38>> Thank you, my friend.
0:40It’s always a pleasure to be here with
0:42you and with guests who help us better
0:44understand these times.
0:46>> Indeed.
0:47And today we have, folks, a very
0:50interesting topic, and one that has
0:52taken over a lot of
0:56the national and international news as
0:59well, because it's a very broad issue
1:02that reaches the whole world.
1:05We are going to talk about the influence
1:08of Big Tech during the electoral process
1:10here in Brazil, and not just that, but
1:13also several other subjects.
1:16with Professor George Valença, who is
1:23a professor in the computer science
1:26department at UFPE, has a PhD from the
1:28UFPE Informatics Center, and is a
1:30consultant for UNESCO.
1:33Professor, a very warm welcome to the
1:35podcast, thank you so much, not only to
1:38you, Vandec, for the invitation. I’m
1:40very happy to speak from the perspective
1:42of academia, right, from the rural
1:43viewpoint.
1:44>> Uh-huh. and for us to bring explanations
1:46for topics that are tricky and current,
1:48right?
1:49I think many people want an explanation
1:51of people trying to parse a phenomenon
1:53that is recent, so perhaps we have many
1:55doubts as well and can't cover
1:56everything, but I think it is essential
1:58that we gain awareness little by little
1:59, because that is how we also demand
2:01answers in some way.
2:03>> Uh-huh.
2:03>> Exactly.
2:05The elections, professor, were not only
2:07contested on the streets.
2:09>> Uh-huh.
2:10>> Right? Nor only on social media.
2:14or on television, or on the radio, with
2:17campaign programs.
2:19The influence of Big Tech, especially
2:23Meta, right, which is the owner of
2:27Facebook, Instagram, and WhatsApp, was
2:31very significant in this election. How
2:35did you evaluate that?
2:38>> I will take a quick historical path for
2:40us to understand that this is not new,
2:42right?
2:44Meta is the company that was called
2:45Facebook, right, until 2021, until the
2:47end of 2021, its name was Facebook.
2:49>> The company is from 2004, so it has,
2:52what? 22 years of existence, right?
2:55And I think the first case we could
2:57highlight is,
3:00I won't necessarily call it influence,
3:01but rather a strong alignment with
3:03political parties, which occurred in the
3:05previous decade, over 10 years ago,
3:07about 10 years ago, right?
3:09It was a scandal that we all know, even
3:11covered in documentaries, called
3:13Cambridge Analytica, which was a
3:15partnership between the former Facebook
3:17—that is, Meta—primarily through
3:18Facebook, its product at the time (not
3:20Instagram yet, but Facebook), and this
3:22data analysis company called Cambridge
3:24Analytica, along with American political
3:26parties, in the case of Donald Trump at
3:28the time.
3:30>> So I think that was a moment when we
3:32began to understand that there could be
3:34a strong alliance between big tech
3:36companies and politics.
3:38Until then, it wasn't a concern for us,
3:40or perhaps we didn't know or have any
3:41idea of the impact it could have. So
3:43what happened at that moment? Let's just
3:45quickly go back to the present, right?
3:47At that time, we had an application
3:49hosted within the Facebook environment,
3:51like a game application, a little game.
3:54And this company, Cambridge Analytica,
3:57in partnership with Meta itself—which,
4:00in this case, was Facebook—gave the data
4:03from this application, and by extension
4:05, from all the users who used it, so
4:08that these companies, together, could
4:10perform a very deep analysis of the data
4:13circulating there, so they could,
4:16together, target what appeared to
4:18Facebook users in favor of the Trump
4:20campaign at the time.
4:22So, it's as if I understood,
4:26it doesn't necessarily have to do with
4:27your game,
4:28>> with what you were playing, but with
4:29your personal data that the game
4:31collected, which is diverse data,
4:32including behavioral data that the game
4:34collected.
4:35And from there, you assemble what we
4:37understand as feeds to favor one party
4:40over the other. So, the data from that
4:42time, which is already, I think, a
4:44decade old—I believe it was 2016, this
4:46Cambridge Analytica case—the posts
4:48appeared with a lot of untruths, which
4:50is a point we've been discussing a lot
4:52ever since, right? A lot of fake news;
4:54we didn't even have that term in our
4:55daily lives back then.
4:57>> So, a lot of untruths and a lot of
4:58targeting in favor of Trump to the
5:00detriment of Hillary at the time. So,
5:02many lies about her, issues that were
5:05sensationalist to bring what? Many
5:07clicks, many clicks in his favor, to her
5:10detriment, and to modulate voter
5:12behavior. And they actually succeeded.
5:14So, I think this is the first moment we
5:16open our eyes and say: "Wow, it really
5:17can exist." And then, ever since then,
5:19we
5:19>> it's a silent advantage, right? You go
5:21there,
5:22>> you don't even notice, and suddenly
5:24you're dealing with a scale of millions
5:25of people, right?
5:26>> Millions. Exactly. Vande, that's a great
5:29observation that we have an ecosystem,
5:31exactly as Gustavo mentioned, Meta
5:32starting from September '21, we call it
5:34Meta, right? Which was precisely based
5:36on other problems that this company had
5:38internally, which I can even list very
5:40quickly right here. Meta is made up of
5:43an ecosystem of products, and they
5:44exchange data with each other.
5:46So, we can have, I’d say, a guarantee—
5:48because we have little transparency, we
5:50don't have a report in my email right
5:52now about Meta—but we know quite well in
5:54academia that WhatsApp, Instagram, and
5:56Facebook are products of the same
5:58company and probably exchange a lot of
5:59information between them, right?
6:02So, an ecosystem of billions of users,
6:03that's no small thing; there are
6:05billions of users there on Instagram.
6:07It's billions, just Instagram alone—
6:09let's forget the others—it's an
6:10ecosystem, a set of billions of users
6:12there with extremely rich data that
6:14allows for what? To deeply analyze our
6:16behavior and then, logically, bring what
6:19we see in the feed—let's talk about
6:21Instagram and Facebook, because we don't
6:23have feeds
6:27that contain information in the
6:28company's favor, not in the user's favor
6:30, which is what we study a lot at Rural
6:31.
6:33In that feed, for example—I'll touch on
6:35children briefly, and I'll get back to
6:37elections in a moment—we don't have
6:38information that is, let's say,
6:40educational or constructive in that feed
6:42. We have information about advertising
6:45, Meta's partners, so that you buy, so
6:47that you click, so that you forward. "
6:49Oh, Jorge, there's no product here, I'm
6:51not even seeing a pair of shoes, but I'm
6:53seeing content here that is considered,
6:55I don't know, very addictive or is
6:56considered content that I will probably
6:58forward to a lot of people." Why? Why
7:00does it also place this content
7:01alongside advertising content? Because
7:03this is content that makes us share a
7:05lot, in other words, nourishing the
7:07network, and it also makes us spend a
7:08lot of time in there. So that's what
7:10they have in their hands, and then,
7:12during elections, it's very valuable,
7:13isn't it? Having this in the
7:14architecture of your product. To them,
7:16that is very valuable, isn't it? They
7:18can do that.
7:19>> I don't know if that already happened in
7:21that Cambridge Analytica episode, around
7:232016.
7:25>> Now we have an issue that was also
7:27widely discussed, which is boosting,
7:29right? So, for example,
7:31>> if I want to boost something to my
7:32friend Zé Gustavo,
7:34>> a Náutico fan, right? Actually, he’s a
7:36die-hard Tricolor fan.
7:39I mean, I boost that that piece of
7:40information, eh, I share this news, I
7:42boost this information.
7:44>> Yes.
7:44>> I don’t boost his response to the same
7:47extent. In other words, this algorithmic
7:49boosting already makes a huge difference
7:51,
7:52>> even though I can protect myself by
7:54saying: I shared the news and I shared
7:55the debunking of the news, isn't that
7:57right, professor? That's it, that's it.
7:59We have boosting that is organic
8:02>> and that is, I’ll even use your
8:03expression, V, that it is veiled
8:05>> and that we have no transparency
8:06regarding it. Unfortunately, we don't.
8:08It's hard for us now, today is Friday,
8:09right? To reach Saturday and say: "Oh,
8:11Jorge mentioned that, I'm going to look
8:12for some Meta reports, but we won't find
8:14them."
8:15Now that the government with the Digital
8:16Act is, well, not quite there, but soon
8:17we’ll return to the Digital Act, it will
8:19be able to demand reports from these
8:20companies.
8:21So, we have boosting that is veiled,
8:23which is their own interest that makes
8:25content reach us through algorithms.
8:28So, their interest, whatever it may be,
8:29commercial interest, possibly political
8:31interest, and having boosting that is
8:32paid,
8:35>> which is what we inject money into,
8:36right?
8:37In the recent elections, if I'm not
8:39mistaken, we can even check, but what I
8:41had read about it, I think about two
8:42weeks ago, is that one candidate, right
8:44, who was Flávio Bolsonaro, he had 20
8:46million, roughly, in investment from the
8:48first to the second round, and Lula had
8:50700,000.
8:53So imagine having, without any political
8:55bias, but imagine having 20 million to
8:56do whatever you want, including
8:58injecting it into this ecosystem, let's
9:00say, right?
9:01An example for boosting. That is a lot
9:03of money for boosting. And that is how
9:05boosting works. If you look at this post
9:07today, we made a video here today of
9:09this interview and they said: "I'm going
9:11to test what Jorge mentioned." I'm going
9:13to boost this video, a reel we created
9:15here today.
9:17You get more reach depending on the
9:21amount you invest. So, if you put in,
9:23for example—let's use random numbers—20
9:25dollars, it says: "I, Meta, right? I,
9:27Meta." I mean, I will let you—I will
9:29make this content reach, and I'll use
9:32arbitrary numbers here. 20 dollars
9:34reaches 100 people. If you put in 50, it
9:36will reach 500 people. If you put in a
9:38million, it will reach.
9:39So, if you have more money, you increase
9:42your reach much more, uh, not
9:44organically, but intentionally. And the
9:46company allows it, right?
9:47>> And, and so, there is, for example, uh,
9:49Vandec and I who come from a newspaper
9:51background,
9:52>> Yes,
9:53>> right? Uh, from a commercial perspective
9:58, doing an organic piece, right?
10:03A piece without advertising ties would
10:06give that company more credibility than
10:10if it were a paid piece, right?
10:14Which the newspaper would frame as an
10:17advertisement, for example, right?
10:21Then a story would come out; on social
10:24media, it's the same way, meaning, is it
10:28possible for organic reach, the one that
10:31isn't paid,
10:35>> to be better or better evaluated than a
10:37paid one, to have more credibility, for
10:40example,
10:41>> more credibility, right?
10:44It's a complex question, and it's a
10:45complex answer for the following reason
10:47. Since we don't have it in Brazil,
10:50internationally, we do.
10:52We have the act or the Digital Services
10:54Act in Europe, which is the DSA, and it
10:56talks a lot about issues related to
10:57algorithmic transparency, the way
10:59algorithms are created.
11:01Since we don't have that in Brazil, we
11:02lack transparency about what the company
11:04is—we're talking about Meta here, but it
11:06could be TikTok, it could be so many
11:08others, right, Telegram, whatever. We
11:10don't have transparency about what the
11:11company prioritizes.
11:14So, if it prioritizes issues that are
11:16sensationalist, issues that are toxic,
11:18like, I don't know, horror content,
11:20violent content, very raw sexual content
11:22, right, I don't know, about female
11:24vulnerability, things related to rape,
11:26something like that, we have no way of
11:28knowing.
11:30So, its organic reach might be based on
11:34values we don't agree with.
11:37And paid reach, sometimes, might even be
11:39more correct. If the government, for
11:41example, the Ministry of Health, wants
11:43to run a campaign with wide reach,
11:44injecting money so that a breast cancer
11:46campaign reaches many people, the value
11:48—here value is subjective, not financial
11:50value, but benefit value—is much more
11:52guaranteed in the paid version, in this
11:54comparison I'm making, than in the
11:56organic one. Because the issue is that
11:58organic reach is based on the company's
12:00interests, on what it can put into its
12:02algorithm, which is a content
12:03recommendation algorithm. That is the
12:05title: recommendation algorithm. It is
12:06artificial intelligence. What it places
12:09there to ensure that content reaches
12:10people is up to the company.
12:13We don't have, for example, a law that
12:15says: "Oh, your recommendation algorithm
12:16—I'm being abstract here—must be based
12:18on values that are human rights,
12:21>> uh, inclusion," that's not the case. So
12:23, sometimes, this content without
12:25advertising or paid boosting, this
12:26organic content, can be just as harmful
12:28or even more harmful. "Professor, we
12:31have the Internet Civil Rights Framework
12:33, we have the LGPD, we have the Digital
12:35Child Statute that you mentioned, we
12:38have the Carolina Dieckmann Law, which
12:40isn't even from your time, but it was
12:42enacted."
12:43>> Thank God it's not from my time.
12:45>> And look, these things still can't
12:48contain what's being done, or are these
12:51frameworks being bypassed?
12:55They are not yet complete enough. The
12:57legal framework we have covers several
12:59things. If we take the Digital Child
13:01Statute, it focuses on children and
13:02adolescents.
13:02>> Anything that isn't about children or
13:04adolescents, we can't interfere with.
13:07>> The Data Protection Law covers issues
13:08related to security, data protection,
13:10and privacy. Anything outside of that,
13:12we can't do anything about, right?
13:14Then there is the 2014 Civil Rights
13:15Framework for the Internet, and it
13:17recently had a ruling, very much to our
13:18benefit as a society, by the Supreme
13:20Court.
13:22This recent ruling took away
13:26from companies this freedom to put
13:29whatever they wanted on these platforms
13:32.
13:33It took away at least that, without them
13:35being punished for it.
13:39So, based on this recent understanding
13:41from the Supreme Court, they now have a
13:43proactive duty to protect the user.
13:47So, it is the due care for the user that
13:49they must guarantee beforehand.
13:52So, for the content circulating there,
13:53you have to guarantee it proactively,
13:55not reactively, not just because a court
13:56decision came out telling you to take it
13:58down. Proactively, you have to assess
14:00risks; you have to verify if that
14:02content is harmful. So, this is a recent
14:04STF ruling, but taken together, these
14:06laws still don't cover everything.
14:08>> Uh-huh.
14:09>> We could have had, for example, the old
14:11Bill 2630, known as the "Fake News Bill
14:13," but it was changed so much over time
14:15that it became a Frankenstein's monster
14:17and was shelved; it doesn't even exist
14:19anymore.
14:20But this Bill 2630, it was known as the
14:22"Fake News Bill," then it started being
14:24called the bill for the regulation of
14:26platforms or social media or social
14:28networks, you know, colloquially, and it
14:31started having provisions that other
14:33laws don't.
14:34First, one I've talked about a lot here
14:35, which is transparency, so it’s not
14:37about opening up the algorithm, opening
14:38the code and saying: "Look, here is my
14:40trade secret, see everything." Isn't
14:41that right? It's about saying: "Look, my
14:43algorithm works based on these
14:45parameters, based on these values, on
14:46these criteria." I use these algorithms
14:48and these techniques. I had an external
14:50audit, not even one done by my own
14:52company, but an external audit that came
14:53here, which I paid for, and the results
14:54of that audit are right here.
14:56Uh, I have reports that include what the
14:58government expects.
15:00Uh, the ones who moderate my content on
15:02the platform are these people, this
15:04group of thousands of people, and there
15:05are also algorithms.
15:07So, these issues, the others don't cover
15:09; the others don't bring them to the
15:10table. That’s why we say we still don't
15:12have day-to-day regulation, right? We
15:13sit down at the pub, ah, there's no
15:14regulation on the internet. Look, in
15:16fact, a completely comprehensive one, we
15:18don't have.
15:19>> Oh professor, uh, for example, getting
15:22back to the issue of, of Meta.
15:24>> Uh-huh.
15:25Uh, in the electoral process
15:29there was on Facebook and WhatsApp, uh,
15:33after the creation of Lula's army, right
15:36, by Janones, which saw gigantic growth
15:40in a very short time.
15:44Uh, even though Flávio Bolsonaro also
15:46had his own channel and several
15:48influencers, including right-wing and
15:50left-wing politicians, had their own
15:52channels.
15:54But nothing compared, compared there to
15:57Janones ', right?
15:59And then what happened? uh
16:03it started to have a
16:08not only did the people from the Lula
16:11campaign start to question Meta for
16:13holding back some posts, right, so as
16:16not to give them reach, for favoring one
16:19side, and disfavoring the other,
16:23>> Just as the Electoral Justice itself did
16:26the same thing, the TCE, the TSE did the
16:28same thing, they questioned the goal a
16:31bit.
16:32Meta came out publicly and said that
16:35they did what they had to do, that they
16:37weren't influencing the race.
16:41>> But there was a council, there’s an
16:44autonomous council there at Meta itself
16:47, that said: "No, the company really is
16:49interfering."
16:51>> How can we have this, this, this
16:53parameter,
16:54>> this discernment, right?
16:55>> This discernment,
16:57>> well, let's see.
16:59I'll, I'll bring up other quick stories
17:01here, look.
17:02>> Let's go back to '25. We're at the end
17:04of '26, right? But if we go to January
17:06of '25, go back a little further. I'll
17:08go to '24 so we can talk about what
17:10context is, what was proven, the
17:12statements, which is evidence. And we
17:14can return to the present, right? If I
17:15lose track of the answer, if I forget
17:17any detail, you remind me.
17:19>> In '24, Meta said the following: "We now
17:21have an inclusion policy." Biden
17:23administration, right?
17:25Biden administration, inclusion policy,
17:27we have teams in here that value
17:29diversity.
17:30We have, I don't know, I'll give simple
17:32examples, say, there are more women on
17:33the teams, there are more trans people
17:35on the teams, we value, I don't know,
17:36human rights more.
17:38And now we have, I don't know, you go
17:40into your Instagram chat and you can set
17:42a theme for the background there, a
17:43background image for your chat.
17:45Vandecia and I put a background that was
17:48for the trans community, it had the
17:50little colors. So they were very
17:51proactive in this process, because the
17:53previous administration had a very
17:54strong focus, let's say, on human rights
17:56.
17:57When it turned to '25, I don't think it
17:59even took long, as if today is November
18:016th, as if it were January 6th, '25
18:03before Trump's inauguration; they
18:05actually changed their discourse
18:06completely and proactively, as soon as
18:08it turned (I think it was January 6th,
18:11if I'm not mistaken), they, Mark
18:12Zuckerberg, right, who is the CEO of
18:14Meta, he gave a speech.
18:16>> that I even used in class, I translated
18:18it, we even used an AI voice for the
18:19students to see in Portuguese, so
18:21everyone could understand.
18:23he gave a speech saying that from then
18:25on they would follow the American laws
18:28that Brazil actually had.
18:30He cites Brazil explicitly.
18:33It's a general address to the world
18:34population, but he says that Brazil has
18:36an obscure court, or it's obscure or
18:38hidden, and that they don't agree with
18:40it and that the laws are American and
18:41that, in partnership with the Trump
18:43administration, he cites: "We will fight
18:45this over-regulation, including European
18:47, on Meta that prevents."
18:49I'm reproducing this, paraphrasing his
18:51speech, but that's what he said, right?
18:52That they prevent freedom of speech,
18:54that they prevent the circulation of
18:55political content, which is something
18:57that will return to the platform. So
18:58they changed completely overnight.
19:00What happened?
19:02Remember when I mentioned the example of
19:03that little chat with a backdrop, an
19:04actual backdrop, an image of the LGBT
19:06community, they changed all the names.
19:08>> Before it was, I don't know, LGBT
19:10community.
19:11of putting cotton candy to try to soften
19:13everything that was contrary to what the
19:15Trump administration, in their view or
19:17in alignment, foresees or advocates or
19:18prioritizes, right?
19:20>> And they went to that famous photo where
19:22Mark Zuckerberg, Jeff Bezos and so on
19:24are, right? Elon Musk, everyone together
19:26there in the pose, right?
19:27So Meta does in fact have a strong
19:29political element. I mentioned the
19:30Cambridge Analytica field, I mentioned
19:31this movement there in January of '25
19:33and also the one from '24 that was with
19:34the Biden administration.
19:36So we know it's a company very strongly
19:38allied with governments to favor itself
19:41. So I don't think it's a company that
19:42we observe throughout its history as
19:44having much neutrality, right? Much, not
19:46the case. The
19:47>> are you saying that big techs, Meta in
19:50particular, is an arm of American
19:52foreign policy. I think it, I won't even
19:55say it flirts, but I think it has many
19:57geopolitical issues in its operations.
19:59>> Uh-huh.
20:00>> This is very present throughout all this
20:01time. If we look at the last decade, in
20:03the year 2000, when I think they were
20:04feeling out, understanding what they
20:06were doing with Facebook, there wasn't
20:08even Instagram or WhatsApp at the time,
20:09right? Instagram was from 2011 and its
20:12popularization came much later.
20:14I think in this last decade, which was
20:16the decade from '10 to '19, right?
20:18They understood what they could do with
20:20our data.
20:21And then there was the Cambridge
20:22Analytica scandal that was revealed and
20:24we had a big change in politics because
20:25of the results.
20:27Look, we don't have a report from Meta,
20:29but there is a series of evidence from
20:31investigative journalism and academia
20:33that reveals that what happened in
20:34politics in 2016 was the result of the
20:36company's intervention at the time,
20:38right?
20:39And it's a company that has acted in
20:40this way ever since.
20:42So, we don't have a report from the
20:44courts here in Brazil saying "yes, there
20:46was interference," but I think the track
20:48record suggests that it is very likely.
20:51>> Now Lula, Lula, just, just a second, my
20:54friend. Lula, in a speech he gave on the
20:5625th, his victory speech, one of his key
20:58points was the following: "we are going
21:00to regulate."
21:01>> Uh-huh. Now, what does "regulating Big
21:04Tech" actually mean? Regulate in what
21:06way, with what instruments? Is it
21:08something within the reach of a
21:10government
21:11>> or does it depend on other factors?
21:14>> Let me just finish; this Meta case, in
21:18particular, shows the vulnerability of
21:24Brazilian democracy in relation to these
21:27Big Tech companies.
21:29>> Uh-huh. Good questions.
21:31Well, let's go.
21:33I saw that speech; I was actually there
21:35for work, I went to three states in the
21:37last 10 days.
21:39And I was even on Paulista Avenue that
21:41day, so I heard the speech there.
21:44I think we have a government that knows
21:45—I'll have to use the Digital Child
21:47Statute example, which for us is a
21:49success.
21:50>> Uh-huh. It has proven to be a success,
21:52even globally. So, the government knows
21:54that with the adjustments that were made
21:55to the Civil Rights Framework for the
21:56Internet, well-made ones, right?
21:58With the Digital Child Statute, which
22:00was a law chewed over and thought out
22:02from '22 to '25, it was 3 years of
22:03maturation; it didn't arise from bile,
22:05it didn't come from an outrage movement
22:07, or a documentary video.
22:09So, it was three years of analysis,
22:10analysis, analysis under that umbrella,
22:11which we’ll talk about in a moment.
22:13I think these moves show an
22:14understanding that the LGPD is catching
22:16on little by little; it hasn't caught on
22:17completely yet, but it's getting there.
22:19>> Uh-huh. Due to the insufficiency that
22:21existed in the ANPD, which is now an
22:22agency.
22:23So the government went there and gave
22:25the ANPD more regulatory power.
22:27Even for the Digital Child Statute, not
22:29just for the LGPD.
22:30So, with these measures and now
22:32increasingly stronger laws, including
22:34the Digital ECA taking hold, forcing
22:35Discord to take things down, etc. I
22:37think the government has a horizon to
22:38say this is not a lawless internet,
22:40right?
22:41A land with laws, federal laws even, not
22:43just state laws. We have many laws in
22:45the United States that are for data
22:47protection, say, from California, for
22:48child protection on a national level,
22:50but there is no general national data
22:52protection law in the United States.
22:53Here we have a federal law, and we are
22:55seeing its success. I think when Lula
22:57talks about this, he comments on this
22:59bill that stalled and was shelved, that
23:01it did, in fact, turn into a
23:02Frankenstein, that was the allegation at
23:04the time, right?
23:07This is probably the path, wait, it has
23:09to be revived, this fake news one, in
23:10heavy quotes, which wasn't just about
23:12falsehoods, so that we cover what these
23:14laws don't cover.
23:16And then we start to have a government
23:17that can demand, as it has been
23:18demanding, and it has had an effect with
23:20the Digital ECA. I think that is our
23:21success story. I'm using the example of
23:23children and adolescents, which is a
23:24Brazilian text, it's a lot of people
23:25>> Uh-huh.
23:26>> who are on the internet, right, all the
23:28people up to 17 years old, babies also
23:29use it now, right, because the friction
23:31is very low, right?
23:32You click, it passes by really quickly
23:33there.
23:35These examples of success make us
23:36understand that internet regulation,
23:38through this new law that is bound to
23:39emerge, we imagine, will be successful
23:41and we will manage it. It’s not just for
23:43show, right? The laws that exist here
23:45haven't proven to be just for show,
23:46where we can't make them exist, right?
23:48In practice, they didn't have teeth.
23:50They do have teeth, right?
23:50>> Uh-huh.
23:52I know you, sir, are not a congressman
23:56or a senator, but we elected, the
24:00Brazilian voters elected an even more
24:04conservative congress.
24:07>> Uh-huh.
24:09>> In relation to the congress that is
24:11ending now in 2026, its legislature,
24:14right?
24:17Do you think, in your experience, it
24:20will be more difficult to pass an
24:23advance in this sense of regulation with
24:26this congress that is there, because the
24:29discourse in '22 and a little before was
24:33already about freedom of speech?
24:38>> right, since the far-right mainly
24:40applies this type of discourse a lot and
24:43I think it will continue a bit along
24:45these lines, do you think it will be
24:47harder or is it still possible to
24:49negotiate?
24:52>> Well, Gustavo, excellent question. I
24:53loved that question for the following
24:54reason, look. There is a thing we
24:55experience a lot in practice, which
24:57isn't an official position in Brazil, I
24:58think it's unofficial, but which exists
24:59a lot in the United States, which are
25:00lobbyists, right?
25:01>> Yes.
25:02>> So, we have people who are actual
25:03lobbyists. It's a job recognized in the
25:05United States, especially by Silicon
25:07Valley, by the countless, very strong
25:09tech companies they have, right?
25:12Excluding India, excluding China, it's
25:13all there, right? The personal data we
25:15have is with these companies in Silicon
25:17Valley.
25:18Here we have unofficial lobbyists, and
25:20beyond them in Brasília, we have a
25:22movement in the broad sense of a lot, a
25:24lot, a lot of lobbying.
25:26I participated in the final advocacy
25:28process for the Digital ECA in Brasília
25:30, so it was a law that went from '22 to
25:32'25, discussed, discussed, many
25:34amendments, amendments to the bill.
25:37And then in the final stretch in August
25:39, I was invited by the Alana Institute
25:41so that we could, also together with
25:43Hugo Mota, right, so that we could go to
25:45the deputies' doors to explain issues of
25:47the law, right?
25:49And we found there, see, a—I think it
25:51was a peculiar scenario, but I'll use
25:53this experience to say how we can deal
25:55with this in practice, right?
25:57I'll take out the agenda that is
25:58children, I'll take out the agenda, I'll
25:59talk about the process in a general way
26:00. We found there a scenario of a lot of
26:03lobbying present. So we would go, for
26:05example, to an office, knocking door to
26:06door. They even say, and on the office
26:08doors it says the following: come in
26:09without knocking.
26:11All offices are like that, right? There
26:13at Congress. So we would enter, ask who
26:15was there, if who was there was an aide
26:17, etc. And sometimes if it was, anyway,
26:19some deputy present, we would explain
26:20the law, sit on the sofa, break down the
26:22law, and so on. There were people, even
26:25the right-wing welcomed us very well,
26:27very well.
26:28Like, ironically we talk about all these
26:30issues, but, you know, they were people
26:32who were very cordial and would say: "
26:33Oh, we haven't even seen it."
26:35When we sat down to discuss, they had
26:37everything highlighted, they had studied
26:39the law, the bill at the time, which was
26:4026, 28.
26:43And we faced very clear issues there,
26:45like when we would enter an office,
26:46there would be a room, and they have
26:48normal meeting rooms, right?
26:50And the door was closed, and we knew
26:51through the grapevine it was a Big Tech
26:54company in there, a representative, a
26:55firm there to dissuade a vote in favor
26:57of the Digital ECA. So, we saw that in
26:59practice. I faced that in practice. We
27:01would leave that office, go to another,
27:02and knock on the door.
27:04So lobbying is very, very present. It’s
27:06not something ethereal that we can say,
27:07"Oh, it doesn't exist, no, lobbying does
27:08exist." I saw it in practice, we saw
27:09people sitting there, right? I went to
27:11Congress the next day to speak upon this
27:14invitation, right, from Hugo Mota, to
27:16talk about the Digital ECA. And then,
27:18before me, Meta spoke.
27:20So it was a vision, I would say, that
27:22was multisectoral. So there were 28
27:23experts who were invited. Daniel Becker
27:25also spoke.
27:27The mother of, uh, Isabela, the girl who
27:30was, right, uh,
27:33>> killed there by her father, was also
27:34present, she spoke, she is also a
27:36right-leaning person, so she spoke from
27:38her point of view, right? So, it was a
27:40very big debate. Why do I think we won?
27:42Because the agenda, which is something I
27:44am bringing as an element for us to see
27:45what we can face now.
27:46>> Yes. Where did we win? How did we win
27:49the Digital ECA?
27:51First, a very well-made bill. It wasn't
27:54a project done in a rush.
27:56There were others that tried to cut in
27:57line, but we strengthened it a lot and
27:59there was a lot of advocacy and a lot of
28:00joint action.
28:02It was a large group, I mentioned 28
28:04experts, but there were other huge ones
28:06there and great people participating in
28:08this fight for this bill, not for other
28:09new and weak ones that popped up in
28:11parallel for people to boost their own
28:13profiles.
28:15So, first, there were many alliances
28:16around the bill, right?
28:19There was a lot of discussion about
28:20everything, a lot of alignment,
28:21including with both sides. Damares, for
28:23example, helped the Digital ECA a lot.
28:25Both sides helped.
28:27Debate with the right, debate with the
28:29left in the hallways, explaining. So
28:31there wasn’t anything we could say—we
28:32spoke about, for example, age
28:34verification by platforms, which is a
28:35super tricky, complex topic.
28:38You say, "Oh, it’s like facial
28:39recognition, it's sending in documents,
28:40" and we explained this so
28:41representatives would be convinced to
28:43approve it, understand it, or propose
28:44something at the last minute, anyway.
28:46So, a matured project, people
28:48coordinating in favor of the project. So
28:51, there was a lot of alliance, a lot of
28:52alliance, a lot of coordination. What
28:54was said, who said it? Let's talk about
28:55something else now, a correct narrative
28:56that had to circulate. We did a lot of
28:58that.
28:59Uh, and the union of both sides and
29:02there was a strong movement, I’d say
29:05national, which I think is also what
29:07saved us. So, people started—
29:09abstracting people, people in general,
29:11right?
29:12Uh, political, non-political, anyway,
29:15the third sector, to defend the Digital
29:17ECA. So, it became such a strong
29:18movement that there was no way for the
29:19government not to approve it.
29:21I mean, this reached, it hit the streets
29:23, right, like, the people.
29:24Yeah,
29:24>> it hit home. And then, who do we credit
29:26this to? This part, to Fel's video, it
29:28raised awareness among people.
29:30>> Now, in this case we’re discussing here
29:32, it’s still a bit distant from the
29:33general public, also because it’s more
29:35abstract, right?
29:35>> It’s more abstract.
29:36>> You aren’t dealing there with people,
29:38right, young people, victims of
29:40>> of some thirst. How, do you think this,
29:42how can we go about bringing this
29:44>> uh more within the reach of the
29:46population?
29:47>> It’s a subject that I
29:48>> it’s that we know it’s difficult, right
29:49?
29:50Because how do we translate this to
29:52explain to people, hey, there’s someone
29:54manipulating there.
29:56I think one of the factors would be not
29:59to let the subject cool down, right?
30:02Because it’s still very hot. Yeah.
30:04>> Uh-huh.
30:05>> I think keeping the subject like this,
30:06not losing, we call it the timing of the
30:08problem, right? For us not to lose it as
30:10if it were the time to sell something.
30:11Oh, this thing went out of style, right
30:13? This thing, this is a gift you gave me
30:15now when you saved up money, I don't
30:16know, a child, right? They go to their
30:17parents: "Oh, but none of my little
30:18friends use it anymore." So we can’t
30:20lose this timing of the trend, the
30:21positive trend, right? And fashion in
30:23general, I'm talking about clothing
30:24trends, music trends. So this momentum
30:26cannot cool down, I completely agree. We
30:28can't let it die down because then we
30:30lose the mobilization. So, just as you
30:32mentioned, Gustavo, that "Janones army"
30:34is an excellent example, right? People
30:36are mobilized, mobilized; we need to
30:37have mobilizations for this too.
30:39And the mobilization will come because
30:41the topic is hot, it's fresh in people's
30:42minds, the feeling is alive there, and
30:44that's how you mobilize based on it.
30:46It's not about manipulating people, but
30:48rather persuasion. You manage to
30:49convince them that they need to work
30:51with you and help you in this process,
30:52for example, and advocate together in
30:54favor of the project, let's say, right?
30:56And I think it has to be very, uh, very
30:59illustrative. So, what if we take the
31:02example of Felk's video once again,
31:04which lasted 50 minutes, right? Why was
31:06it very clear and had so much appeal?
31:07And that's what J also managed to do.
31:09Anyway, the elections also brought a lot
31:10of this, because there was an example of
31:12what child sexualization was. Uh-huh.
31:14>> There was an example of what child
31:16manipulation and co-optation was. Here's
31:18the video and here is the guy's name.
31:20Uh, another bill we're not talking about
31:22here, which hasn't been advancing in
31:24Brazil for a year—I think it stopped in
31:26December—is the AI bill. We don't have
31:29AI regulation in Brazil, we don't.
31:31And we have the EU AI Act, which has
31:34been the AI law in Europe since '24.
31:38We don't have it in Brazil.
31:40I discuss this in the classroom like,
31:41look, when we have it, or rather, there
31:43were many public hearings throughout all
31:45of '25, there were hearings to debate AI
31:47and the environment, AI and copyright,
31:48and anyway, but we don't have it.
31:52How could you, for example, bring the
31:54public closer by saying: "Look, when we
31:56don't have AI regulation, we will have a
31:58company that offers, for example, an AI
32:00boyfriend, and this will affect your
32:02child, because they might date an AI,
32:04right?"
32:05So, I think these very clear examples of
32:07appeal—not sensationalist appeal, but
32:09rather the harm is clear. I think these
32:11examples bring people closer, right?
32:13>> A more concrete example, right?
32:14>> More concrete. It has to be clear. I
32:16think it can't be.
32:16>> I think there was a very strong case, it
32:19was at, uh, Folha de São Paulo, perhaps
32:21others.
32:21>> Uh-huh.
32:22>> Where they made, let's say, art, like a
32:25poster or banner with, uh, Flávio 22 and
32:28Lula 13.
32:30The Flávio 22 one they did without any
32:32issues, right, Meta,
32:33>> but the Lula 13 one, no.
32:35Later they argued it was a technical
32:37glitch, but look, it's still a point to
32:39discuss, right? I mean, you got one
32:40right, you do one and then the other, I
32:42mean, perfect. And then,
32:43>> this, this is AI, this is a case that's
32:46>> easy for anyone to understand, right?
32:48Because in this case it was Flávio and
32:50Lula, but later on it could be anyone,
32:52right?
32:53George, Gustavo, Vandec, right? It's
32:56just that.
32:58And they are interesting because, for
33:00example, something that struck us: they
33:02are malleable and have a lot of room for
33:05maneuvering.
33:07Maybe in Canada, sir, correct me if I'm
33:10wrong, but Canada banned, uh, required
33:12that if news were published by the press
33:14, Meta had to pay royalties.
33:17And what did they do? They banned the
33:19news.
33:19>> Uh-huh.
33:20>> I mean, I even thought, wow, the
33:22newspapers are going to get revenue
33:23because all day people are posting news
33:25articles.
33:27On social media, they think the
33:28following: no, I lost, done, zeroed the
33:30law,
33:30>> zeroed it. It's true, isn't it?
33:32>> You kill by starvation, right?
33:34>> You kill by starvation.
33:35>> Yeah, so it's a, it's a
33:36>> it's a way for you to dodge it, right?
33:37They get around it very well.
33:38>> They get around it.
33:40>> We have, for example, safety locks in
33:42ChatGPT.
33:43I work a lot with the production of
33:44content for children and adolescents,
33:45right? And there are many agendas we
33:47discuss, articles we write, studies that
33:49touch on sensitive topics. We don't talk
33:52, for example, about, as a rule, we
33:53aren't going to discuss, I don't know,
33:55gamification of education, we are going
33:56to discuss the manipulation of education
33:58, we're going to discuss apps for the "
34:00nudification" of children and
34:01adolescents, which is the production of
34:03sexually explicit material of children.
34:05Our topics are heavy, right? We dampen
34:07the mood when we go to a conference.
34:10They are all very...We received an award
34:11last week for an article in Minas Gerais
34:13about, colloquially speaking, right?
34:16Adolescents who date AI.
34:17>> So, the topics are heavy. When I put a
34:19text like this into GPT,
34:20>> they are heavy, but they are extremely
34:21current, right?
34:22>> They are extremely current.
34:23>> because it happens, it happens a lot
34:25nowadays,
34:25>> a lot. This is very frequent. That's why
34:27we have to debate it while the
34:28phenomenon isn't so widespread yet,
34:30right? Kind of to contain it, right?
34:32>> Academia has this role. Our group
34:33prioritizes this a lot. What happens
34:35when you try to post it? In my
34:36day-to-day, I go to make a post about
34:38this, I say: "Look, we have, we had an
34:40award-winning article about, uh,
34:42notification platforms, where you put an
34:44image of a teenage girl with clothes on
34:46and the platform generates a nude image
34:48of her."
34:49"Make a summary for me to post on
34:50Instagram, a news piece about this."
34:52It says: "Look, unfortunately," it puts
34:54it in red, "unfortunately this content
34:55violates our policies."
34:57>> Uh-huh.
34:57>> It's an AI and it's an AI that has locks
34:59. It doesn't stop me. It's actually
35:01appropriate content. It's a professor
35:02wanting to share academic work. I can't
35:04summarize it. "Look, my article talks
35:06about all of this.""Help me quickly
35:07summarize this here to make a story." It
35:09doesn't lock. Why didn't it lock Flávio
35:1222?
35:12>> Uh-huh.
35:13>> So we can implement locks.
35:16>> This thing you mentioned about going to
35:18Congress, to every deputy and senator's
35:20office, what entity were you
35:21representing?
35:22It was a group that was
35:24>> UFRPE. UFRPE.
35:25>> I was invited as a professor
35:27>> but together with other
35:28>> together with other entities, very much
35:30in partnership with our research group,
35:31which is the ASPAS group, the word ASPAS
35:32, right?
35:33>> which is an acronym, a sigla for
35:35Sociotechnical Aspects in Software
35:36Platforms. So we investigate the part,
35:39let's say, of algorithms, code, etc.,
35:40but always with a view to understanding
35:42impacts on users, risks, harms, and
35:44benefits too.
35:46We were invited and we work a lot in our
35:47research group with the third sector,
35:49and with the government as well. We have
35:51a very long-standing partnership with
35:52the Court of Audit here, 20 researchers
35:54who work at the court right nearby,
35:56right? I'm very close to the court.
35:58>> Yes, but we also work a lot, we have
35:59funding from the third sector from the
36:01Alana Institute, I work with UNESCO,
36:02right, with UN agencies.
36:04Eh, and there we also worked a lot with
36:06the third sector. We would go to these
36:08doors, uh, knocking with the third
36:11sector, which I think does this advocacy
36:13process very well, of championing these
36:16causes.
36:17>> So, this mobilization is worth keeping
36:18up, right?
36:19>> It's worth keeping up, it's feasible,
36:20they receive you.
36:22Professor, speaking specifically about
36:25your work there at the Rural,
36:28>> and based on what we've already
36:30discussed here, you always bring up the
36:33issue of children and youth.
36:37Besides being a sensitive and extremely
36:40current topic, I wanted you to give an
36:44overview of how
36:47>> this work that you carry out there at
36:49the Rural is done.
36:51Talking a little about the importance
36:53and the impact that this
36:55>> can cause. Like, it's a curiosity of
36:58mine,
36:59>> how do you manage to burst the bubble,
37:02>> right? And bring this to society in a
37:05clearer way.
37:06>> Yes. Excellent. I also loved the
37:08expression "bursting the bubble," which
37:09is something we use a lot in the
37:11classroom and in our courses, right?
37:12There is a particularity; the computing
37:14department—I've been a professor in the
37:16department since 2013—is a major
37:17department. So, I've been there for 13
37:19years.
37:20The department has grown a lot over time
37:22, so we have more than 50 professors;
37:24there are many people, it's a large
37:25department.
37:27I have colleagues across Brazil who have
37:29a computing department with 15 people to
37:31handle a huge number of students. Since
37:33we have more than 50, we can dedicate
37:35ourselves to the areas we understand
37:37best, like most, and have more
37:38motivation for. And what we do as a rule
37:40in the classroom, we do in research and
37:41vice-versa. One thing feeds back into
37:43the other, right? So what I research, I
37:45also teach in the classroom, in the
37:46courses, and vice-versa.
37:49Our research group works on these issues
37:51that are linked to two major groups.
37:54One group is linked to age range, which
37:56is children and adolescents. So, if you
37:58look, you deal with platforms, we deal
38:00with them, but what do you look at on
38:02the platform? Issues related to children
38:03and adolescents. So, the vision is more
38:05oriented toward age range, which could
38:07be the elderly, because that’s an age
38:08range too, right? Or young adults, and
38:10we also work with a focus that is
38:12oriented toward race and gender. Then,
38:15in that other focus, we study non-binary
38:16trans people. And the impact of facial
38:19recognition technology on them.
38:21This umbrella of children and
38:22adolescents in our group is much larger
38:24because there are so many problems
38:25within it, so many problems. How do we
38:27burst the bubble? First, we deal with
38:29topics that are very current. We avoid
38:31looking in the rearview mirror. Not a
38:33distant rearview mirror, no. A rearview
38:35mirror like
38:35>> recent,
38:37>> lately, looking at a social network
38:38isn't even that interesting for us.
38:41We try to look at phenomena that are
38:42emerging. So with UNESCO, what did we
38:44investigate throughout this whole year?
38:47AI Companions, which is a brand new
38:48English term from late last year, which
38:50we translate as AI companions, they are
38:52affective AI companions, there are over
38:54100 platforms focused on affective
38:56relationships that can be romantic.
38:58We also investigated AI Toys with UNESCO
39:00, which are toys with Generative AI,
39:02which are also countless. We analyzed 28
39:04, generated a report, the Ministry of
39:06Justice jumped on it, and it was very
39:08successful. These two studies have been
39:10very, very successful. We even spoke
39:12with a radio station in Russia, from
39:13Russia, about the topic. So, it even had
39:14international repercussions.
39:17And we dealt with notification apps, so
39:19these are forward-looking topics, topics
39:21that people don't know about yet.
39:23So it is a foresight vision, it is what
39:25we call speculative design. We are
39:27trying to foresee phenomena that haven't
39:29spread yet. How do we burst the bubble?
39:31They have appeal, so everyone wants to
39:33know.
39:33They are very palatable. We show screens
39:35, we show data, we have testimonials.
39:38Uh, and we also make them the result of
39:41strategic partnerships. So look, not to
39:44diminish it, but we don't have a
39:45partnership that is, I don't know,
39:46between universities.
39:47We have one with UNESCO, one with the
39:49Alana Institute, which is an institution
39:50funded by Itaú, which has a lot of money
39:52and can do a lot, and one with the NPD.
39:56And then I also have a lot of openness
39:57to talk to the media, so that is also a
39:58way to burst the bubble. I was on
40:00Provoca just now, with Marcelo Tas, this
40:02week.
40:02Uh, so that's a way to talk, I speak to
40:04a lot of written media, a lot of
40:06podcasts, it's a way for us to reach the
40:08general public, because if it's just
40:09academia with articles and conferences,
40:11it won't move forward, it won't.
40:13>> Now, there is also the people's side,
40:15because there are many companies today
40:17that work alongside these big techs,
40:18local companies, right?
40:20>> A lot.
40:22>> So there is a fear that this could cause
40:25some noise, some alteration, or a
40:27reduction, right, in investments.
40:31So it's even hard to find people who
40:33talk about this subject,
40:34>> we looked, it's not very common either.
40:36>> That's true.
40:37>> Because you can expose yourself,
40:38>> right? You have to,
40:40>> you can, you have to be very cautious.
40:41>> You have to be very cautious
40:42>> because sometimes you're, I was in Belém
40:44in May at a debate with Meta; at the
40:46table, it was me and a Meta
40:47representative.
40:49Then there was I think someone from
40:50SaferNet, which is an organization,
40:52right, representing the UK here in terms
40:53of digital protection and citizenship. I
40:56think there were four or five of us, a
40:57student, a teenager. So I have to be
40:59very cautious because in the debates I
41:00do, I might even be on the same side as
41:01a Big Tech company.
41:02>> Uh-huh.
41:04>> Sometimes questions are for both of us,
41:05or I have to be cordial, careful with my
41:07words, but in principle, we don't
41:08prioritize partnerships with—we're not
41:10closed off to it, but partnerships with
41:12companies where we lose autonomy.
41:14>> For sure.
41:15>> Or with political parties, right? I get
41:16invited, but I think I could lose my
41:18autonomy and my word could lose its
41:20weight, right? Because it might look
41:21like I represent a party. "Oh, George
41:23made a plan for the party, an IT plan
41:25for the party, a computing plan, a
41:27digital transformation plan for the
41:29party.""Look, George is working with the
41:31party," so that's not something I want,
41:33>> in principle, because then we lose this
41:35autonomy that we have.
41:36>> Of course.
41:37>> And my word is all I have, right?
41:38Exactly.
41:40>> Uh, Professor, do you think it's
41:43possible to create a more democratic
41:48internet and
41:51and a safer one?
41:53>> Uh-huh.
41:54>> Especially for children and teenagers
41:57nowadays.
41:59>> I think so. I see the glass as half full
42:01, and it's not based on, it's not a
42:02Pollyanna view of things, right? I think
42:04we actually have a pillar that is
42:06regulation, which in Brazil is enviable
42:09.
42:09In Latin America, it's already by far
42:11one of the best, but in Brazil, for the
42:13first time, Europe is inviting us to
42:15listen to us about digital success.
42:17Instead of us going there to drink from
42:19the fountain and, I don't know, hearing
42:20in Switzerland what they're doing, in
42:21France, no, it's the opposite. They now
42:23pause to listen to us, right? There was
42:25a delegation, I wasn't there, but
42:26several of my partners went, right? From
42:29the Ministry of Justice, the Alana
42:30Institute, from various institutions to
42:32talk about Brazil's success in
42:34regulation. So, we have regulation that
42:36is doing just fine, we are moving in the
42:38right direction, and we have an agency,
42:39because otherwise the regulation would
42:41have no teeth. The ANPD has over 500
42:43people nowadays, and there is hiring
42:44coming up in January with a new civil
42:46service exam. So, you are growing it
42:48with top-tier professionals; professors
42:50are very frequently invited to fill
42:51positions at the ANPD.
42:55This makes it so the law can actually
42:57stick, because you have oversight and
42:58specific training for that law.
43:00So, it’s like the people at the ANPD who
43:03are responsible for the Digital ECA, for
43:05the LGPD, know what manipulative design
43:08is, know what algorithmic moderation is
43:10, right, content moderation.
43:12And we have more and more, I think,
43:14awareness.
43:15Obviously, we want more. We want basic
43:17education to teach a 14-year-old to know
43:19what Discord is, how you protect
43:21yourself from, I don't know, an AI
43:22companion. What should a little kid do
43:25here? My nephew, for example, let me
43:27open a parenthesis. My nephew, since he
43:29was little, we might have had some
43:30bad-taste jokes at home, a bad habit,
43:32right?
43:34Maybe we let out some expression once
43:36that was wrong, and he went, "Uncle," I
43:38don't know if it was to me, "Uncle,
43:40that's bullying," let's say, I don't
43:42know, calling someone fat.
43:45"Uncle, that's bullying." I think he was
43:46, I don't know, about 7 years old.
43:48Because he was a child of a generation,
43:50he's from 2013, that learned that
43:51calling someone fat can be bullying.
43:53Uh-huh.
43:53>> He said it naturally, didn't make an
43:54annoying face, nor was he too happy, "
43:55Uncle, that's bullying."
43:56>> It's not activism, he had already
43:58internalized it. So that is awareness.
44:00So we hope it comes from the teachers,
44:02who are so dedicated, such fighters. We
44:04want these people, who already have a
44:06Herculean task, a huge effort there, to
44:08understand all of this, which is a
44:10challenge. So I'll see how the Ministry
44:12of Education can help. Uh-huh.
44:13>> Because the awareness has to be theirs
44:15for it to reach the students.
44:17And the general population, the families
44:19. So this literacy part is very
44:20important and there must be a change in
44:22IT practice.
44:24It's no use having regulation, etc., if
44:25IT, which is the big tech companies and
44:27who I train—I train computer science
44:28people, right—doesn't change. I get down
44:30to practice and say: "Look, Jorge, I
44:31taught a class on manipulative design, I
44:32don't care about that, I only care about
44:34the money." You can't do that. We have
44:35to have classroom discussions about it,
44:37we discuss Afrofuturism, a futurism that
44:39looks at and lets a group take the lead
44:41that has never been the protagonist.
44:44So people have to leave the classroom
44:46saying: "I understand the importance of
44:47defending people against facial
44:48recognition with bias."
44:51So you have to change this practice by
44:52raising awareness among those entering
44:54this market and you have to make this
44:55market respect the law in its day-to-day
44:57operations, because a fine might just be
44:58a tickle.
45:00>> And, and you can see from everything he
45:04said here, the importance of incentives
45:07for science and education in
45:09universities, right?
45:12Because I imagine the desperation
45:17>> that was the 4 years of the Bolsonaro
45:19administration, which took away so much
45:21incentive from
45:23>> education and the universities, right,
45:26professor?
45:27>> Yes, we have to have the most diverse
45:29types of funding, right? Because we
45:32mainly have to fund the students. A
45:33professor, at the end of the day, has
45:35their position and salary, they are a
45:36civil servant, right? But they also need
45:38support, and it’s not just about valuing
45:41their work and their working hours.
45:43Sometimes they need resources to buy a
45:45computer. I'm talking about my area.
45:46Think of the various fields of knowledge
45:48, right? Resources to buy a computer. If
45:50it’s a biology professor, to buy, I
45:51don't know, components for their
45:52laboratory, or chemistry, something like
45:54that.
45:54So, you need resources for that, but you
45:56really need it to keep the students.
45:59If you don't have a scholarship, the
46:00people with you need a lot of motivation
46:02to keep fighting by your side, otherwise
46:04it's just volunteer work, right?
46:07Then if a request from the company comes
46:08and a request from your work, let's say
46:10, with UNESCO, the company's wins.
46:11>> Of course,
46:12>> because everything has a limit. Even my
46:14activism as a student here has a limit.
46:16My company is demanding things from me
46:17and they are the ones who pay me, right
46:18? Then you have to say: "No, I have a
46:19scholarship from, for example, I have a
46:21scholarship from the Alano Institute, I
46:22have a scholarship from the Audit Court
46:23, they pay me.""Oh, okay.""So I have
46:25hours to work on the project and to
46:27study something I want." And that is my
46:29thesis, that is my dissertation, right?
46:31>> So, when the government injects money
46:32into research, you come out the other
46:34side, and you can solve things, right?
46:36Yes.
46:37>> Because otherwise, you’d have to host a
46:38conference with your own resources.
46:39Imagine, I just attended a national
46:41conference, but it was in Quixadá, which
46:43is part of bringing these congresses to
46:45the interior,
46:46>> in Ceará.
46:47>> In Ceará. And I thought it was beautiful
46:49that the organizers—the women—I heard
46:51behind the scenes that they paid for
46:52some things out of their own pockets
46:54because the funding sometimes isn’t
46:55enough. They wanted to put on a show and
46:57make the conference stunning, and it was
46:58great that the interior and the
46:59Northeast showed what they’re made of,
47:01right?
47:02But there was something—I don’t know
47:03what it was—there was this wonderful
47:04popsicle cart just sitting there. Maybe
47:05that was the cart they bought, or maybe
47:06it was the gift bags, maybe it was the
47:08bags, but they definitely spent their
47:09own money. So even when the government
47:11funds it, you still need sponsors, you
47:14know?
47:15>> Yeah, it’s important, folks.
47:18We’ve had a very good, spectacular
47:20conversation with the professor, and I’d
47:23like to thank you very much for your
47:25presence, Professor, and say that 9.1 is
47:28always open, right?
47:30If you have any new work or something
47:31you’d like to promote there at the rural
47:33university, you know?
47:35In your field of computing, right?
47:38This podcast isn't exclusively about
47:40politics. We're talking about politics
47:43at this exact moment because it was the
47:48news of the day, but it’s open to all
47:52other topics. You are always very
47:54welcome.
47:55>> I appreciate it, guys. It was a pleasure
47:56, and it's great that we’re able to
47:58break out of our bubbles as academics,
47:59just like you said.
48:00>> Exactly. Perfect.
48:01>> So, thank you very much. Thank you, guys
48:02. I hope we have other opportunities to
48:04debate, and I hope many people listen,
48:06like, and share the idea, so more people
48:08become aware of these tricky topics.
48:10That’s true.
48:12Folks, I’d like to thank Vandec for
48:14being here once again.
48:16We are wrapping up today’s 9.1 podcast.
48:19We’ll be back next week and will be
48:23joined by Minister André de Paula here
48:27at our round table.
48:29See you, and until next week.

About this transcript

This page contains the full transcript of Podcast 9.1 | Big Techs e a influência nas eleições 2026 com o Prof. George Valença | 06/11/2026 by TV Guararapes Oficial, generated from the public captions YouTube serves with the video. The transcript has 9,450 words across 1,527 segments, with the original timestamps preserved so you can click any line to jump to that moment in the embedded player.

What you can do with it

Use the transcript to take notes, quote the speaker, build a study guide, generate a summary with ChatGPT or Claude via the YouTube Summary tool, or export it as a timed subtitle file with YouTube to SRT. You can also re-open it in the transcriber to choose a caption track available on the video.

Free YouTube transcript tool

YouTube2Text is a free YouTube transcript generator — no signup, no daily limit. Paste any YouTube link and get the full transcript instantly, with timestamps, click-to-jump, caption language selection, AI prompts for ChatGPT, Claude, and Gemini, and exports to TXT, SRT, VTT, or Markdown.