DiPLab’s Antonio Casilli Featured in the Culture Pages of Magazine L’Espresso

DiPLab co-director Antonio Casilli is featured in the culture pages of the Italian magazine L’Espresso, where he discusses the process that led to the release of his award-winning documentary In the Belly of AI, which he will present at the Festival del Pensare Contemporaneo in Piacenza, Italy on September 13, 2025. The 75-minute documentary, co-written with Julien Goetz and directed by Henri Poulain, exposes the hidden human costs behind artificial intelligence systems—from data centers consuming massive environmental resources to underpaid “data workers” in the Global South who process disturbing content to train algorithms, often developing psychological trauma in the process.

Casilli’s appearance at this international philosophy festival, themed “Vite Svelate” (Unveiled Lives), aligns perfectly with his documentary’s mission to unveil the invisible labor powering AI systems. The film has earned critical recognition including the Outstanding Excellence at the Documentaries Without Borders festival the documentary has been shortlisting for the Melbourne documentary Film Festival.

Find here the pdf version in Italian and the html version in English of Casilli’s article.

EspressoDownload

Last June, when Meta announced the appointment of Alexandr Wang to head its new artificial superintelligence program, the trade press hailed yet another enfant prodige of Silicon Valley. At only 28 years old, he has become one of the most influential faces in the race toward General Artificial Intelligence. A textbook American parable: brilliant MIT student, billion-dollar deals, rise to the top of the tech industry.

Behind this glossy narrative, the reality is another: Wang is first and foremost a cofounder of Scale AI, and his entry into Meta coincided with the nearly $15 billion investment Zuckerberg’s behemoth made in his company. Founded only a few years ago, the startup that turned him into a star in the industry relies on a business that is anything but glamorous: employing millions of data workers who for paltry sums enter information, tag images, transcribe text, and filter sensitive content. These are the riders of artificial intelligence: precarious, underpaid jobs that fuel the technology of the future.

While investors and the public are being dazzled by the rhetoric of superintelligence, behind the scenes is this industrial army that continues to work for Scale AI and hundreds of other such companies. The truth is that artificial intelligence, no matter how sophisticated it may appear, is still massively dependent on the most precarious human labor on the planet. An untold number of digital workers, estimated by the World Bank to be well over one hundred million globally.

For years, the media have documented the working conditions of these “click proletarians.” In January 2023, two months after the launch of ChatGPT, a Time magazine investigation brought to light the existence of real digital sweatshops in Africa, where workers were paid as little as a dollar and a half an hour to “train” OpenAI’s artificial intelligence.

Also supporting Time’s reporting was the research of academics like myself, who had been engaged in field investigations for years to document this reality. At that time I was already working with director Henri Poulain on what would later become our new documentary, In the Belly of Artificial Intelligence (In the Belly of AI, Federation Studios, 2025). My research activity has long been intertwined with the production of television investigations on the relationship between technology and work. This time we decided to shine a spotlight on a burgeoning phenomenon: the use, to make artificial intelligence work, of the most vulnerable people. From the victims of armed conflicts, to the millions of unemployed in the poorest continents, to low-income workers even in rich countries. These are not just companies like Scale AI, but international chains of exploitation in the digital economy.

Uma Rani, an economist at the International Labor Organization in Geneva, explains this with disarming clarity during one of the first interviews in our documentary. Every time ChatGPT answers a question, a Tesla car avoids an obstacle, or Instagram recognizes a face in a photo, there is a stream of human labor running across the planet behind that automation. It’s not just code and algorithms: it’s sweat, toil and exploitation in countries where a few cents can make the difference between surviving and not making ends meet.

But this interview was just the beginning. During months of filming we met dozens of people: inmates in Finnish prisons who for two euros a day train specialized intelligences, Ukrainian refugees reduced to tagging digital images to support their children, Indian migrants tagging images for a few cents an hour. But it was especially the slum workers in Nairobi, Kenya, who struck us the most. Not only because of the extremely exploitative conditions, but because of the devastating psychological consequences of their work. Many of them train the artificial intelligence of large multinational corporations not to generate illicit content: rape, torture, murder, abuse. This kind of activity, also called “moderation,” represents the darker side of artificial intelligence training. Even those who moderate content on Facebook are actually teaching automated systems to block violent or offensive photos and text. Those who perform these tasks often develop post-traumatic stress disorder after spending months viewing and cataloging gruesome content.

In the face of these testimonies, we bumped into the wall of silence from Big Tech. What is the response of AI development companies in the face of all this? Virtually nothing. Apart from a few timid attempts to give themselves codes of ethics that have no legal standing, their main job seems to be to silence critical voices. The realization of In the Belly of Artificial Intelligence, more than in our other projects, has been a minefield: interviews with experts and witnesses canceled at the last minute, lack of approvals from institutions and government departments, intimidation by lawyers linked to large corporations and law enforcement.

The most surreal episode happened in Nairobi, where we were interviewing some artificial intelligence trainers. While we were filming, a group of police officers raided the venue and pressured the crew to turn off the cameras. Kenya is in a particularly delicate position in dealing with tech giants. The country has become a crucial hub for these kinds of digital services, hosting untold numbers of data workers who work for global AI giants. This economic dependence makes the Kenyan government particularly receptive to pressure from multinational corporations. A few months after the police raid during our shooting, President William Ruto himself publicly intervened in lawsuits against companies that exploit workers at starvation wages. He announced an amendment to prevent future lawsuits against these companies: ‘We changed the law so no one can take you to court anymore.’”

What we have realized during these years of work is that this army of data workers is not “invisible labor,” as it is often called. It is actively hidden by tech companies that use intimidation, legal hurdles and media manipulation to construct a storytelling that makes people like Alexander Wang the sole heroes of the digital revolution.

But the real heroes are others. Ordinary people, without fame or recognition: the women and men who today are finally organizing in unions and associations to demand decent working conditions. It is they who, thanks in part to documentaries like ours, can now hope that their voices will finally be heard. Because behind every intelligent algorithm there is always human intelligence. And that intelligence deserves respect, dignity and justice.