Democracy Now! Audio - Democracy Now! 2026-08-27 Thursday

Episode Date: August 27, 2026

Democracy Now! Thursday, August 27, 2026

From New York, this is Democracy Now. We protect our children from so many daily dangers. But we can't adequately protect them from harms meta hides and lies about, or from designs that are built to suck our children in. With this settlement, that is changing. Meta, the parent company of Facebook and Instagram, has agreed to pay up to $17 billion in a landmark settlement after the company was sued by states for designing platforms,

that are addictive and harmful to children. We'll get a response. Then, as Bill Gates sounds the alarm about the risks and dangers of artificial intelligence, we'll look at how AI companies are using a largely hidden global workforce to train their models. We'll air excerpts of the film in the belly of AI. All kinds of artificial intelligence needs humans behind them. Our army of workers. There is reports from the World Bank in which they actually argue that there are around 150 million data workers

and 430 million data workers around the world. The workers are invisibleized and the workers are invisibleized. What are they trying to hide? We'll speak to Ambakaq of the AI Institute and Antonio Kazili, the author of Waiting for Robarts, and Julian Posada, author of Platform Extractivism. All that and more, coming up. Welcome to Democracy Now, Democracy Now.org, The War and Peace Report. I'm Nareen Sheikh.

More than a 350 people are dead and over 1,300 remain missing after a massive wall of water, mud, and debris crashed through a Himalayan River Valley on the border of Nepal and Tibet. It's the region's worst disaster since a devastating earthquake over a decade ago. The mudslides swept through the Lendejola River Valley just after 8.30 a.m. on Wednesday, washing away homes and burying entire villages. Water levels along the river at one point rose by as much as 27 feet over half an hour. This is Sabina Tamang, whose husband remains missing.

In our last conversation at 7.30, he said he was sending two kilos of water. Olmets and two kilos of chocolates for the children. He also sent some instant noodles for them, and he said he would be coming soon, but he didn't. All I want is for my husband to come back to me safely in one piece. I don't want anything else. Hundreds of foreign tourists are among the missing. The mudslide washed away at least 19 bridges, complicating rescue efforts. The disaster was triggered by a 2,000-foot-long chunk of ice that broke away from a glacier high in the Himalayas. Climate scientists warned the region is warming much faster than the global average, leading to the rapid retreat of glaciers. Nepal has

lost nearly a third of its total ice volume over the last three decades. The Pentagon said it carried out strikes against alleged drug boats in the Caribbean Sea on Tuesday, killing four people. The latest attack brings the number of those killed, according to the Pentagon's own figures, to 227, across 68 strikes that began almost a year ago. Once again, the Pentagon provided no evidence the boat was carrying drugs. Amnesty International has condemned the strikes as extrajudicial killings, a form of murder, saying they amount to crimes under international law. The chief of Iran's nuclear program says inspectors with the International Atomic Energy Agency will not be allowed to visit nuclear sites in Iran that were bombed by the U.S. and Israel.

On Wednesday, Mohamed Islami, the head of the Atomic Energy Organization of Iran, criticized the IAEA for failing to condemn the strikes and said the UN agency was being used by Iran's enemies as an intelligence gathering tool. The IAEA is pressing to visit the damage sites under pressure from the United States in Israel. The reason for that pressure is that they want to see what their military operation has done to those sites. On Wednesday, President Trump told Al Jazeera he has no time schedule for when he expects to return to the negotiating table with Iran, adding, quote, I'm not in a hurry. Ukraine resumed drone attacks on Russia on Wednesday, just hours after CIA director John

Ratcliffe departed Moscow following an unannounced meeting with Russian intelligence officials. Following Ratcliffe's visit, a wave of Ukrainian strikes killed 12 people inside Russia and in the Russian-occupied Donbass region of Ukraine. Meanwhile, Russia's military targeted nine Ukrainian cities overnight, with officials in Ukraine reporting hundreds of ballistic missiles and drone attacks. The barrage damaged residential blocks, a school, and a medical facility in Kiev, and killed a 78-year-old woman in Zaporizia. The attacks came, as Russia's foreign ministry has warned it could target British military assets after the UK supplied Ukraine with long-range missile technology. We have repeatedly warned that the response to Ukrainian strikes using British weapons on Russian territory could target any British military facilities and equipment, whether in Ukraine or beyond, those responsible for war crimes, including those committed against. our country's civilian population will be punished in accordance with their actions. Meta has reached a landmark settlement with attorneys general from 47 states, the District of

Columbia, and U.S. territories, over claims it intentionally designed its social media platforms to addict children. Under the settlement, Meta must pay more than $17 billion in penalties over a decade. It also agreed to block overnight access to its sites while setting a two-hour limit for anyone under 18 and will silence most notifications during school hours. This is Victoria Hinks reacting to the settlement. Her daughter, Alexandra Hinks, died by suicide at age 16. I think as long as they enforce it properly, and I'm glad that the judge asked, how are you going to actually do this? Because sometimes it's just a lot of talk. And they need to actually have independent people come in. I think it's great. I think this is three times now, with three,

trial in L.A., they were found liable. In New Mexico, they were given a huge fine, and then now this, they settled. We'll have more on META's landmark $17 billion settlement agreement later in the broadcast. Connecticut Senator Chris Murphy is calling on the Trump administration to halt immigration raids in the city of Danbury, following reports that ICE has been arresting parents at school bus stops. Murphy wrote on social media, quote, ICE waited until the kids were on the bus, when they are dropped off at the end of the day, they are alone, scared to death they can't find their parents, unquote. Senator Murphy joined other elected officials in Danbury Wednesday at a rally of immigrant rights groups demanding accountability.

Congress should not authorize a single additional dollar for ICE while this illegality continues. The advocacy group, Danbury Unites for Immigrants, counted 25 people detained Wednesday, 16 on Monday and at least nine on Tuesday. The group says detentions occurred at day laborer gathering spots and a Latino grocery store and that two of its legal observers were shoved by officers with one hospitalized with injuries from pepper spray. Meanwhile, five-year-old Liam Tadeo and his father, Victor Martinez Nieto, who were detained by ICE and Austin, Texas earlier this month, were deported to Mexico on Tuesday. According to the fact, family's lawyer, Liam's father had begged officers to let his son go as they were arrested on their

way to soccer practice. His father had wanted his son released to his mother, but that option wasn't provided to him. A judge has blocked Minnesota's bid to extradite Christian Castro, an ICE agent accused of firing his gun and wounding a Venezuelan man, Julio Cesar Sosa Sassali, in the leg. Castro remains in custody in Texas. Under Texas law, Castro could be released from jail today. Texas governor Greg Abbott has said he will not abide by Minnesota's extradition request. This comes as Alex Prettie's family calls for accountability from the Trump administration seven months after he was fatally shot by Customs and Border Patrol agents in Minneapolis.

Prettie's parents said they found out their son had been shot while watching the coverage of the shooting on the news. His mother said she recognized her son's jacket and sunglasses from the footage. Two CBP agents have been charged for use of force, but they have not been charged for the killing. Alex Preti's parents spoke to ABC's Good Morning America. My shock is that he was murdered, and no one has given us the complete truth or anyone has been held accountable. And it's been seven months. There was more effort in vilifying him than... pursuing justice, investigating this.

A federal appeals court has declined to dismiss charges against Democratic Congressmember La Monica McIver, who visited the notorious ICE facility Delaney Hall as part of routine congressional oversight. She was charged with interfering with an arrest in May 2025 outside Delaney Hall. She's the only Congress member indicted by Trump's Justice Department. Earlier this year, Democracy now spoke to Congress member McIver about. about the charges. The process is the punishment.

This is what they're doing. They're doing this with your taxpayer dollars, prosecuting a member of Congress for doing their job. It doesn't cost them anything because they're using taxpayer dollars to do so. But I think what it does is it inserts fear and other leaders to step up and hold the administration accountable. It's intimidation. It's bullying.

And they're just using me as a tool and a prop to do so. And the Trump administration has imposed sanctions on the government. UK-based direct action group, Palestine Action, and two other organizations, designating them as a specially designated global terrorists. Palestine actions members have broken into factories that produce arms for Israel's military where they sabotaged equipment. The U.S. sanctions follow a UK appeals court ruling last June that found the British government acted lawfully when it banned Palestine Action under the Terrorism Act.

Palestine action co-founder, Huda Amori, responded in a statement saying, quote, The fact that Trump is now taking inspiration from Britain's repression of the movement for Palestinian freedom exposes just how dangerous this ban is and should be a wake-up call to anyone who cares about free speech and civil liberties, she said. And those are some of the headlines. This is Democracy Now, Democracy Now.org, the War and Peace Report. I'm Narmine Sheikh. We begin today looking at the landmark meta settlement. On Wednesday, the parent company of Facebook and Instagram agreed to pay up to $17 billion to settle a lawsuit brought by 29 states led by California that accused the company of designing its platforms to be addictive to children and harmful to their mental and physical health. The social media giant will initially pay about $12 billion and an additional $5 billion if Snap, TikTok and YouTube.

to reach similar agreements. Meta also agreed to a number of restrictions for users under the age of 18, including limiting their access to Meta to two hours a day and silencing notifications during school hours. Meta will also block their access to its platforms from midnight to 6 a.m. This is Victoria Hinks, mother of Alexandra Hinks, who died by suicide at the age of 16, reacting to the settlement. I think as long as they enforce it properly, and I'm glad that the judge asked, how are you going to actually do this? Because sometimes it's just a lot of talk. And they need to actually have independent people come in. I think it's great.

I think this is three times now. With trial in L.A., they were found liable. In New Mexico, they were given a huge fine. And then now this, they've settled. A key witness in the landmark trial warned meta-safety changes fall short of protecting young news. users. Whistleblower and ex-Facebook employee Arturo Bechard said, quote, the limitations that are in the agreement are the equivalent of saying, well, you can smoke as many cigarettes as you can in two hours a day. It doesn't make the cigarettes any safer. The meta case is being called a bellwether

federal trial, potentially marking an inflection point for an industry that has largely escaped regulatory efforts. The meta settlement also raises larger questions about the role of government in regulating powerful technology companies, especially in the age of rapidly expanding artificial intelligence or AI. Well, for more, we're joined now by Amba Kak, co-executive director of AI Now Institute and former senior advisor on AI at the Federal Trade Commission. Amba, welcome to Democracy Now. If you could just begin by responding to the settlement, it's significance. It's significant and it's welcome, particularly for an industry that isn't very used to being held to account for its impacts on the public.

And $17 billion, roughly, across 10 years, you know, it's a big number. It's certainly a big number compared to the fact that the Trump DOJ was able to muster up a poultry 400 million from TikTok just about a week ago in a similar child privacy case, right? But I think, you know, we've reached that point and the mother's testimony speaks to this, is fines no matter the volume, particularly for companies for whom even $17 billion is, you know, a couple of days of what they're spending over 10 years on AI infrastructure, right? So for them, it's a drop in the ocean. But maybe more importantly, if we look at what these, you know, meta is agreeing to in this settlement, it's design changes, it's design fixes. But we're not

really getting at the root cause of why do we have infinite scroll or why are these platforms being architected in ways that are designed to addict teenagers? And the answer is the business model. It's the business model of surveillance advertising where it is optimization of engagement and attention at any cost. And I think if we don't get at those root cause fixes, we're going to be stuck, you know, playing whack-a-mole once the harms have already metastasized. Well, just to go back to, you know, it is $17 billion, but in fact, it's at the moment $12 billion and the remaining five is contingent on what meta, YouTube and TikTok do. What do you think the chances are that they would also agree to them?

Look, I mean, firstly, this is $17 billion. It's over 10 years. But to their point on whether industry will follow suit, I think at this point, the evidence of the harm involved in this case of mental health impacts are unavoidable for the industry. We've reached the point where this is the least they can do to prevent harms that have metastasided and have already caused, you know, suicide. So I think this isn't, this doesn't seem like a place that they're going to negotiate. I think for these firms, they're always looking to the next frontier and where can they

eke out the maximum sort of space. Their recklessness in this case, I think they've already sort of been clamped down on. you explain what is the what is section 230 of the communications decency act and how has it been used in these cases so you know if firms have a great deal of immunity for user generated content but what's happened is that they often use this as a shield to say we're just a neutral platform but i think what these cases really bring out that there are these mechanisms of optimization optimization of our attention and these design features that are

that are very much transforming these platforms from much more than just a kind of neutral vessel for content, but certainly shaping our attention flows in ways that they get to architect and whether that's to buy things or for the kind of political opinions we shape, they're playing a much more fulsome role in shaping information flows.

And so they use, I think that is why section 230 and the like have come. come back into focus because people really are questioning the basis of these immunities in the first place. So you talked about the root cause. I mean, it is true this problem or this fact of this, the infinite scroll. What are the ways in which those are being the fact of this root cause of infinite scroll, among other things, how are those being addressed? And is there something, is there the role of government here? Who can regulate or in any sense alter what is now practically universal across social media.

Precisely, I think what this case tells us is that we cannot be relying on fines or certainly even on litigation after harms have already transpired. What we need are rules of the road. And we need rules that go to the root cause of the business model. Now, like I said, in the case of social media, that's the surveillance advertising business model. How do we actually, you know, shut that engine down so that we're not stuck dealing with the symptoms of that engine. And I think these are the lessons we need to be taking to AI right now.

Not in five years, not in 10 years. This is the moment to set the rules of the road and really protect and put public safety and public health first. So just tell us, you know, you've been advising governments, including the U.S. government on artificial intelligence. What are you telling them and what kind of responses are you receiving? So honestly, Narmeen, it has been quite incredible to see how the landscape, even in my engagements with governments and regulators all over the world, it's been a complete 180. I think even a year ago, the orientation was very much, you know, if you bring up regulation, your anti-innovation, you're a Luddite, you're seen as anti-AI, and, you know, politicians were sort of figuring their way around that

and the landmines around being labeled as anti-progress or anti-innovation. But cut to today, I think in prominent part because of data center activism across the country, but also a general feeling among the public that AI seems very anti-people. It seems like it might be coming for our jobs. It might be coming for our kids' health. I think the fact that public opinion has sort of made itself heard and is very concerned about these technologies and their impacts has also meant that I think the public, once again, has the listening year of the political class.

question now is, you know, what do we do with it? So let's talk about those data centers and more broadly the infrastructure that's required for artificial intelligence. You co-authored a piece in the Wall Street Journal in November in which you talked about the cost of the infrastructure and what even these massive tech companies, Google, Microsoft, Amazon, etc., that have powered AI, the amount that they would have to generate far exceeding what they actually have. Right. So the industry, the AI industry, the AI industry at this point has chosen to go all in on an extraordinarily capital intensive trajectory, right? Bigger is better into infinity. And the question really is sort of who is going to foot

the bill. And what we argue and have been arguing is that actually the government in this case, and particularly the White House, the Trump White House, has taken an orientation that goes beyond just sort of reassuring industry that they will not come for them with regulation. it's a step further. They're acting not just as customer, but really is underwriter for this trajectory. They're doing this through public procurement contracts of, you know, large numbers. They're doing this through the commitment of federal lands for data centers,

basically, you know, laying out public lands for the use of AI data centers. They're actually doing the bidding of AI companies abroad to get contracts for these companies in different parts of the world. So I think there's a way in which, and maybe most importantly, right, every time there are jitters about is there a bubble, is there not a bubble, is any of this going to pan out? You have the president himself standing alongside, you know, Sam Orkman and Larry Ellison, just sort of reassuring everyone, but really reassuring the market that this administration is going to stand behind these companies no matter what. So I'd like to just turn to someone who's warning about the potential dangers of artificial intelligence and its trajectory. Last night, during an interview with CNN's Anderson Cooper, Microsoft co-founder Bill Gates, warned about the many societal risks of unregulated artificial intelligence development. This is what he said.

Over the last year, these AIs have gotten dramatically more powerful, even faster than I expected. And they're now capable of causing cyber attack risk, bioterrorism risk, psychosocial risk. And I have to say I'm kind of shocked that the exact criteria that we review these models with and the actions we take to minimize the harms are really completely missing. And the broad dialogue, we get all of society talking about the tradeoffs here, that's just not taking place. And so, you know, we've got to minimize these harms through careful review and criteria. And we have to accelerate the good stuff or else I know we could have a complete backlash that AI would suffer from. So, Amba, if you could respond to Bill Gates and what he said, I mean, he was initially an AI enthusiast and the importance of someone in his position making the statement.

Look, I'm glad that Bill Gates is coming around to the fact that maybe AI is not such a great thing for the public. But I think there's some amount of fatigue, you know, and I don't think I'm speaking personally, even among the public, with sort of looking to tech messiahs as potentially having the answers or being the experts on questions of whether technology is and how to shape technology so that it is eventually socially beneficial for the public. And I think that question of who is going to sit as experts. in this question of AI harms and what to do about them, I think is a really important political question in this moment because you asked me about what it's like speaking to governments in this moment. I think the one thing that they do grapple with is a sense of insecurity

and confusion about perhaps not knowing enough. And this is manufactured because you have an army of lobbyists that are continually reinforcing that they will sit expert in solving problems that they've created. So this is no shade on Bill Gates, but I think the era of looking to, you know, tech CEOs or tech luminaries to give us answers for what are fundamentally social and political questions, I think, is gone. Okay. So before we conclude, I just want to ask you earlier this year, India hosted, became the first developing country to host the AI Impact Summit. You were there if you could talk about the significance of India's, India hosting this. And then how India is kind of positioning itself as a third alternative, not China, not the US, but India. Look, I think the India question is in some ways much broader than just India.

I think it's the question of if we sitting here in the United States are worried about corporate consolidation in these few big tech firms and our digital infrastructures being controlled by the tech overlords, then that has reached the point of being an existential crisis for the rest of the world that has seen over the last five years what that looks like when you have core digital infrastructure controlled by companies that are eventually beholden to erratic demands from the White House. And that is absolutely a crisis. Now, what to do about it, I think where the India summit was interesting is that instead of submitting to the notion that we're all in this race against China

and there's just, you know, we all have to be all in on this one race or else. I think there was an effort by India and many other countries, I think to try to eke out the space to say, are there other ways? Are there other ways to build these technologies that not only have the kind of harmful environmental and other impacts that the scale paradigm does, but potentially can also eke out some space for, you know,

alternative innovation ecosystems and strengthen local economies. So I think the headline was, look to your own population and see what they need rather than being sort of pulled by the terms of this geopolitical great power race, which frankly is unwinnable on material terms anyway for most of the world. Thank you so much, Amber Kuk, co-executive director of the AI Now Institute and former senior advisor on AI at the Federal Trade Commission. She's advising governments and government bodies on regulating. AI. Coming up, the hidden global data workforce, we'll speak to Professor Antonio Kazili, author of Waiting for Robots. Stay with us. Endless Alternate by Dominique Gerard Bernard. This is Democracy Now, Democracy Now.org, the War and Peace Report. I'm Narmine Sheikh. One of the most discussed aspects of how

artificial intelligence will transform society is how it will impact jobs and labor. On Wednesday, Bill Gates published a nearly 6,000-word essay in which he warned that society is not ready for the transformation and upheaval AI will cause. In the essay, Gates calls AI's impact on the labor force one of the three primary risks of the technology, saying many jobs will disappear forever and that, quote, AI will either be the greatest equalizer ever invented or the worst source of injustice. But what gets much less discussion is how artificial intelligence has already transformed work globally. Our next guest is sociologist Antonio Casili, author of Waiting for Robots, the Hired Hands of Automation. He writes, quote,

human workers make automation possible. Recognizing the value of their contributions to these digital infrastructures is the first step towards worker reparation. and rebalancing the current power dynamics. He's also co-writer of the film in the belly of AI. This is the film's trailer. AI can be a transformational tool in our fight against climate change. We are really optimistic about the potential for AI to help scientists cure, prevent, and manage all diseases in this century.

Some robots will eat some jobs, and some robots will kill humans. But when you look at the trend as a whole, I think it's going to be incredibly positive. for Inhabited. All kinds of artificial intelligence needs humans behind them. Our army of workers. There is reports from the World Bank

in which they actually argue that they're around 150 million data workers and 430 million data workers around the world. The work is invisibleized and the workers are invisibleized. What are they trying to hide? Can you take a book?

Please. Yes, for sure. A cloud is not a cloud. It is concrete, right? So it's cement. Carbon footprint, carbon emission of the data center is huge because most of the electricity that they use come from fossil fuels. There is one AI company which says that it usually

remove people from poverty. In your opinion, is that true? It's manipulation. It's just a bunch line they use to attract people who have people who have. attract people who are qualified. What does this AI development? How does it really help our society and the development process? Is it something that we really need?

That was the trailer for the film in the belly of AI, directed by Henri Poulin and co-written by our next guest. Antonio Cazili joins us from Paris. He's professor at the Institute Polytechnique de Paris and author of Waiting for Robots. Professor Casili, welcome to democracy. now. Thank you for joining us. If you could begin by talking about these data workers, workers around the world whom you've met, who are automating AI. Yes, data workers are today's factory workers insofar as AI has become like a driving force for so many industries. and they are necessary to produce what we use every day,

you know, chapboards and models. Now, we are not talking about tech workers, engineers, product managers, designers, these highly paid and highly recognized professionals. We are talking about persons who are hidden, and I insist on the fact that they are not invisible. They are purposely hidden by the companies that produce AI. because they are the secret ingredient of AI itself. They train it, and this is a verb that we are now more familiar with,

because we know that, I don't know, chat GPT has been trained for a number of years by persons who were literally feeding them data and enriching this data. But they are also these data workers, those who verify that, for instance, the results of the same ChargerPT are in line with what we are looking for as users. And sometimes they are even those who, you know, operate maintenance, perform maintenance live, meaning that sometimes they replace the machine itself and so they, to an extent, impersonate AI. So they do all these things and they are really poorly paid. We have been following and working with them for a number of years, for six years now already.

And we have been conducting service and interviews and observations in more than 30 countries and interviewed more than 4,000 of these workers in places like, I don't know, Venezuela or Kenya or Madagascar and so on. And they are really everywhere. They are also in the U.S. and Europe, also in high-income countries. But of course, given the very low wages that they have as a compensation, these kind of jobs, which are not jobs, these are more tasks that they perform in a more or less formal way.

They are more interesting to people who are living in low-income countries. And if you could explain Professor Kazili, I mean, one of the things that's so shocking, because indeed these workers are hidden, if not invisible, as you say. Just the sheer number of workers there are. I mean, the World Bank report found that there were anywhere between 130 and over 400 million workers, gig workers. I'm not sure how many of those gig workers are data workers. But if you could explain also, I mean, the very curious, because you've been to Madagascar, you've been to Kenya and various other places, including Finland and Bulgaria,

the way in which these workers are compensated. In some case, as you've said, in Madagascar, they're not paid in money, but in sacks of rice or sugar. Sure, yes. This is something which is appalling, which is unfortunately very common. First of all, let's focus on the estimates that the World Bank has been producing for a number of years because they started, you know, conducting this kind of service and trying to estimate the number of people who perform online gig work. We're not talking about gig workers like, I don't know, Uber drivers

or delivery persons, you know, delivery workers. We are talking about people who perform online tasks. So they are, you know, facing a computer and they perform small tasks like, I don't know, they have to describe an image or they have to transcribe a sentence. And all these is fragmented in small bits of information, that information that we call data,

and they are put in databases, and then the models, AI models, use this information to learn. And this is why we talk about machine learning. So this is necessary. But because it is necessary, it has to be this kind of work process, has to be spread all over literally hundreds of millions of persons. And now the World Bank in 2015 already estimated this type of workers in the tens of millions.

And now in 2023, they updated their estimated to hundreds. Now, they are very vague as to how many exactly. So this is why they talk about something halfway between 100. and something, 100 odds million and 400, which would represent something between 4% and 12% of the global workforce. Now, personally, I am a bit suspicious of this estimates, but one thing that I can say, given the fact that with our other colleagues, we've been trying ourselves to estimate this work as all around the world in several countries, we have noticed that over the years, these estimates have been going up and up. So we're talking about probably 160 million persons in 2021, and now the estimates is bigger.

And this is also something that we observe because we are literally traveling the word, trying to give voice to this hidden workforce. And we notice that more and more persons are available to speak with us. and also organizing to protect their rights. And you're right, we've been to places like Madagascar and Kenya. And in some cases, we encounter people who are paid a few cents. This is the case of Venezuela, where at a moment, even the factor of earning a few dollars, especially because it was in dollars or in cryptocurrency, but not in Bolivares,

which is the local currency, which was depreciating. from one day to the other. So this was interesting, meaning this was an appealing type of job, let's say. And in other countries, like for instance, Madagascar, where we spent, well, a few months over several field works, we basically encountered at least one instance of a company that provided the opportunity for workers to connect to a... a platform to work and to perform tasks on platform and they were paid in sugar, rice, beans and other foodstuffs. So when we say that they are paid peanuts, well, that's a metaphor,

clearly, but in this case, it was not a metaphor. It was something which was clear, visible, to the extent that they also had some kind of a meter at the very entry of this company to explain to the workers how much they would earn in terms of food for each level of performance, let's say. Well, Professor Ghazili, let's listen to the data workers precisely in their own voice. Many data workers are tasked with reviewing highly disturbing content almost exclusively. This is former Kenyan data worker Faustine Makira in Kenya in a clip from the documentary in the belly of AI. A warning to our listeners and viewers. It contains discussions of sexual abuse.

Well, at first, I didn't see it as something that was hard, but as time went by, I realized that I used to have nightmares, especially when I've watched a lot of murder cases or rape cases, especially if it involves minors, children. when I have watched a lot of corpses during the day I would have difficulty sleeping I developed anxiety I could not go out I could not be

where people were many in a crowd I also felt like I was not being myself because I was this an outgoing person but most of the time because I was scared to go out I would just be alone and it kind of led me to depression and that is why I decided to quit. Today there are still traces of the PTSD

or the things that I'm going through that is why I'm still going for therapies. About my anxiety, most of the time you will find me indoors so yeah I would say they're still detress. Can you take a break, please? My co-workers, they did suffer the same symptoms. just depended like with you. I have a colleague who is now a friend. He's battling with insomnia. I have a friend who is also battling with anxiety. So it just depends on how you

as a person was affected because you are affected differently. But yes, we have very many, I have very many colleagues and friends who are also battling with PTSD in different ways. Why did not? Didn't you use the actual name of the clients and the companies you worked for? Because they made a sign an undisclosure form, which ties me to not say some things. Immediately I say them and they go out there. I will be jailed for more than 10 years. Yeah.

So that's a clip from in the belly of AI. directed by Henri Poulin. Professor Kazili, if you could just explain, you know, it's not entirely clear. What do these data workers do with this bad content? You've called it something like machine unlearning. And then we want to get also to the way in which these organizers against these workers against impossible conditions

are, in fact, trying to organize against their working conditions. Sure. Well, I call this machine unlearning because sometimes these workers have the terrible task that consists in flagging information. They do not want the models, the AI, to reproduce. Like they don't want AI to create disturbing images or illegal content. They don't want AI to perform actions that are concerned. considered, again, illegal or anti-social. So they are somehow responsible tasked with the difficult task of performing this kind of teaching, which is a negative teaching. And when we say

moderation, content moderation in particular, the first thing we think about is social media. And we think that this is not connected to AI. Well, to the extent that AI exists on social media too, and Meta has its own AI, and TikTok has its own AI. It is true that moderators can work both for social media

and for AI, but it's not only censorship. They are teaching the machine not to see some things, to remove them from the database. And this is something that goes beyond the simple flagging of problematic content because we have met people who perform this kind of machine unlearning

also outside social media. I don't know. Somebody who has to train an AI that is used in surgery has to go through so many gruesome and bloody images that can be traumatizing and can create the same type of PTSD of post-traumatic stress disorder. Somebody who has to train, I don't know, driving car. They have to go through, for instance, car accidents that can be extremely morbid and extremely problematic to watch itself. And they can develop this kind of traumatic mental health issues. So it is something that is connected to the very process of teaching the machine to do something or not to do something. And this is also something that, you know, as a, as a, as a talk on these persons. In this case, and again, I want to thank Faustin that I have met again in Kenya

a few months ago, I mean, for her testimony, but she is someone who paid an important, a serious, a severe personal price. Some other people were even less fortunate, less fortunate, because some other people died because of these committed to suicide or died in mysterious ways that are somehow connected to their activity. And this is something that we are documenting in many countries, in Kenya, in particular. And so we are also trying to work with the families to face this kind of serious problems. So Professor Kazili, could you explain why are workers made to sign non-disclosure agreements? and in some instances, as documented in the film,

they aren't even able to speak to one another about the content that they're reviewing. Why is that? Well, there are many reasons. The first one would be that they should not, according to the tech companies, talk about the kind of model that they are training because this could constitute an issue

in terms of intellectual property. For instance, if this company is developing a new, model which is not on the market, they don't want the competition to know that. The second reason would be, yes, to actually invisibilize, as we say in a rather pompous way in academia, meaning to hide this worker. So in order to hide them, the first thing is to, you know, well, shut them out literally. And so that they, even if they are asked, they deny that they are actually working to produce AI performing this kind of. kind of data work. And one final reasons is that, again, according to the designers of these

AI models, if workers speak among themselves and they start discussing the kind of tasks they perform, this could introduce specific biases, meaning that they could, for instance, coordinate. This is the technical reason. But there is also a very, you know, reason that is connected to, you know, breaking the solidarity between workers. If workers are segmented, fragmented, they are separated from each other, and they do not speak to each other. They do not understand, although they do not reach the awareness of the fact that they actually are doing the same kind of work, they are one class of workers, and these class consciousness does not develop. Professor Kazili, we'll have to leave it there.

Thank you for joining us, a professor at the Institute Polytechnique de Paris, and author of Waiting for Robots, the Hired Hands of Automation. He also co-wrote the documentary in the belly of AI, directed by Henri Poulin. Coming up, Julian Posada, author of the forthcoming book, Platform Extractivism, Data Work and the People Powering Artificial Intelligence. Stay with us. High Fly by the legendary pianist and composer Randy Weston. We continue our discussion of the hidden data. data workforce powering artificial intelligence with Julian Posada, author of the forthcoming book,

Platform Extractivism, Data Work and the People Powering Artificial Intelligence, which is a deep look at data work and data workers in Venezuela. Julian Posada is assistant professor of American Studies at Yale University and co-director of its computing culture and society certificate program. Julian, welcome to Democracy Now. Now, if you could, we just heard from Professor Antonio Kazili about data workers, you focused principally on Venezuela. If you could tell us what you found among data workers there.

Yes. So when I started this project, I actually wanted to focus on Latin America when I grew up. And the first thing that I found when I started this research is that most of the workers at the time when I started around 2020 were located specifically in Venezuela. So part of the project in the book is about why. why Venezuela contains so many of the workers that were powering, I call it the book Powering Artificial Intelligence at the time. I found three reasons for that.

The first reason is, of course, the economic conditions of the countries. This is 2014. The economy of Venezuela has been in a crisis, hyperinflation, high unemployment rates. And adding to that, you have then the COVID pandemic starting, increasing unemployment rates with the lockdowns. So it was really, really a dire situation in the country around seven, six years ago. The second reason, however, why Venezuela became so prominent in data work is the infrastructure of the country,

which was developed by the government. So, for example, people have access to computers that were created in Venezuela, built in Venezuela factories during the Hugo Chavez years, that then were retained by families. And those same computers were the ones used for data work during the pandemic. And then the third reason, which is also an interesting one, is the composition of families in Venezuela. What I found is that from the qualitative interviews, not a single worker was alone. Workers really depending on their families. And during the pandemic, when a lot of people lost their employment, it was just one breadwinner in the house.

They did a worker. And the entire household was supporting that work through many activities, feeding the worker, doing house children, so on. but the family really revolted around data work and then the neighbors as well. If you lost access to electricity, if you lost access to something, your neighbors would be there to provide support. So this is why I think Venezuela became such a prominent case for data work in Latin America and in the world during the pandemic. Again, the economic crisis, the infrastructure and the composition of families. And could you tell us how, what were the salaries like for these workers, what were the working conditions like, and who were they working for?

So salaries were either very low. Some platforms, a platform remote task would pay workers per hour, and it would be 50 cents of a dollar per hour. All their platforms, the majority of them, would pay per task. So it would be a form of piecework. Workers would earn around $2 to $5 per week because of the hyperinflation then. Those few dollars were important for their income, right? And the interesting part is that some workers, and this is tied to the question about, the number of workers and their location. Some workers would use VPNs to master IP addresses,

pretend they're in countries like here in the United States, and then get paid more for the same amount of work spreading not to be in Venezuela. So it's really difficult to, one, estimate who they are, where they are, but also their pay rate really varies across tasks and platforms. And how do we, do they have any idea for what major tech company they're working? Is like, does Open AI or Anthropic, do they, how do they employ these? workers effectively through, how does it work? So a company would usually approach outsourcing company. Let's say, for example, you have your AI startup, you approach one of the outsourcing companies,

for example, Scale AI, which is a company behind Remote Tasks. Then with that company, you would then create a project and then outsource it to one of their platforms, in this case, remote tasks, and they would have programs, in this case, for Venezuelan workers. then from the perspective of the worker, you would just see the tasks. In most cases, you don't know who they're working for, what tasks you're doing for, or what the AI that you're doing is in reality. I asked some workers if they had any guesses.

For example, one worker was telling me, I think I worked for a task related to the military complex because I have to tag roofs and bridges in some desertic place somewhere, but I have no idea where this is, and I have no idea what the company or what the AI training is actually for. So there was really complete invisibility in that regard. The workers would not know who their employers were and what even the AI they were helping develop is.

And if you could talk a little bit also about how the workers are monitored and how their work is quantified. Yeah, so like in any case of gig work, the algorithm here is the manager. And there are some cases in which workers, like in the case of Kenyan workers or workers in Madagascar, for example, where they're working for companies inside.

They go to an office, like a call center. They're working from a computer. They have their boss around, and they go to a single location. That's what we in the academia called the business processing outsourcing centers, BPO's. My work in Venezuela was primarily focused on platforms. So the best way to think about it is think of the Uber of data annotation. So you will log into your computer to a platform on your browser, and then you would access the tasks and then do the tasks.

And in this case, the algorithms would be the ones, the manager algorithms, the ones actually controlling the workflow. You would have a timer. They would know how long you've been in each task. They would try to guess the accuracy of your tasks. They would present the same tasks twice or three times. So if you're doing different things, the algorithm will think you're spamming. So I'm going to ban you from this task now.

So it really was an algorithm put in place to control the workforce and ensure that they were doing the job that the clients wanted them to do. And what was the thing, were you able to speak to a lot of these workers? Yes, I did. And what was the most frequently reported thing that they told you? Well, one of the three things is that really depending on the tasks they were doing, they really varied in terms of payment because of the investments. So, for example, there was this task called Peace Mode,

in which workers would have to label images in households, let's say an image of a living room. You have to label the couch as a couch, mirror as a mirror, and so on. And workers would then be paid bonus for this task that were really important for the companies and the firms. In this task, they were very repetitive, most of the cases. And at the time, so this is pre-LLMs, right? So at the time, there was a lot of investments on computer vision, investments, for example, to develop facial recognition algorithms or object identification algorithms.

So this is why Venezuela, in which the workers, many of them, didn't have knowledge of the English language, were able to tag these objects and train this image recognition algorithms. Then post-2023, when the investments started to focus on large language models, this is when ChagipT came out, then a lot of that work was moved from Venezuela to places with a colonial legacy with English-speaking countries, so with Philippines or India, in which many of these workers were required to speak English or knowing the English language, then were tasked with any test that we do with language, in the English language in this case.

And at the same time, the situation in Venezuela improved a little bit. Oil prices started to increase again. And a lot of these workers then stopped doing data work and went back to their regular activities. Thank you very much, Professor Julianne Posada, assistant professor of American Studies, and co-director of the Certificate in Computing, Culture and Society at Yale University. He's the author of the forthcoming book, Platform Extractivism, Data Work, and the People-Powering Artificial Intelligence. And that does it for today's show. Amy Goodman will be in Middlebury, Vermont, Friday night, for a screening of the film about Democracy Now called Steal the Story, Please, with the Oscar-nominated directors.

Then on to Madison, Wisconsin on Labor Day weekend, check our website at DemocracyNow.org for details. Democracy Now is produced with Mike Burke, Dina Gazzar, Anjali Kamath, Messiah Rhodes, Nicole Salazar, Maria Tarasena, John Hamilton, Sarah Nasir, Churina Nadura, Sam Alcoff, T. Marie Astudio, Robbie Karen, Hani Masood, and Diego Ramos. Our executive director is Julie Crosby, and very special thanks to Becca Staley, John. Randolph, Paul Powell, Mike DiPilippo, Miguel Negera, Hugh Grant, Carl Maxer, Dennis Moynihan, David Prude, Dennis McCormick, Matt Ely, Anna Osbeck, Emily Anderson, Dante Toretti, and Buffy St. Marie Hernandez. I'm Nermin Sheikh. Thanks so much for joining us for another edition of Democracy Now.