The Turing Point. Newsletter 42AI Impact and Governance #2
EDITORIAL : Reinvesting the Digital World, Renewing the Time for Dialogue
“Who should decide where to draw the line and how to weigh the pros and cons? CEOs of companies or democratically chosen governments?”
Reasoning through arguments against taking AI safety seriously — Yoshua Bengio
Technological accelerationism, the climate crisis, multiple wars… These challenges now find a common denominator in Artificial Intelligence. And according to the AI giants, the future will inevitably be written with their technology. But is this truly the case? Are we condemned to an artificial intelligence that merely assists humans in what they already do best at their worst, before eventually surpassing them, simply?
This ideological narrative permeates the media landscape, saturating it. Yet, one hand would almost suffice to count its bards: OpenAI, Anthropic, Meta, Google, X… The possibility of a different trajectory for technological progress is thereby denied, deemed impossible, and thus unthinkable. The task is not to let ourselves be discouraged. These issues are immense, but collective: it is possible to act at our scale.
It is time to renew dialogue among all actors in this complex ecosystem that represents AI. To create meeting spaces, workshops that put tools in the hands of women, minorities; to reflect on their design and limitations with the youngest and the oldest. To make them more accessible, more ethical, more transparent, more inclusive.
To the giants of large language models (LLMs), alternatives already exist. Aleph Alpha, Lumo, AMI, smaller models accessible via open-source through Hugging Face — the GitHub of AI models — or French-led projects by CNRS. Every day new initiatives emerge, each more open than the next. It is also in this effort that 42AI mobilizes to organize events, boot camps, and conferences, the first of which was held on May 22nd (watch the replay here, in French). We believe in valuing all points of view so as not to leave technology solely in the hands of entrepreneurs and engineers.
We must grasp these questions because we all inhabit, to varying degrees, the digital world. We cannot merely be tenants of it.
Article — The AI Bubble: Has It Burst or Not Yet?
For nearly a year now, numerous voices have diagnosed a speculative bubble around the generative AI industry. Sam Altman, CEO of OpenAI, even conceded as much in August 2025. Achieving profitability in generative AI already requires a high-wire act during “normal” times. But many factors could make this exercise even more perilous. Between energy, climate, and geopolitical crises, welcome to the era of “polycrisis,” which further weakens the AI bubble. Let’s illustrate this with a focus on chips manufactured in Taiwan.
Article written by Benjamin Karas, currently retraining at 42 Paris and co-head of “The Turing Point” newsletter.
In Hal’s Eye: The magic behind AI
Generative AI often feels like magic: it seems to “understand” us, its answers are “brilliant,” its generated videos are stunning… But behind every model lie errors, hallucinations, and, crucially, humans. Many humans.
We easily imagine the brilliant developer making AI models perform miracles. We often forget the flip side: the data annotators.
Their mission is to label images, videos, or 3D environments by identifying elements within them: trees, trains, cats, etc. This processed data then becomes usable for training AI models. If ChatGPT generates you a magnificent image of the Paris Metro, it’s because it was trained on thousands of Paris Metro images tagged with the label “Paris Metro.” And just like that, the magic happens.
Today, the job of a data annotator is becoming increasingly complex, requiring more skills: video labelling, medical imaging, LiDAR 3D renders, and more. Despite this, it remains subject to numerous controversies. The wages of these workers fall far short of Silicon Valley developers’. Their working conditions are sometimes degrading from a human standpoint. The latest scandal: Kenyan data annotators were required to watch intimate scenes of people using Meta Ray-Ban glasses as part of their work. Social inequalities are not about to disappear “as if by magic,” even with AI.
Article — Algorithmic Video Surveillance During the Paris Olympic Games
The French state experimented with Algorithmic Video Surveillance (AVS) during the 2024 Paris Olympics. Despite strong controversies, approval was granted to continue the experiment by expanding its scope. This provides an opportunity to evaluate the first phase during the Games: what constraints were imposed for the trial? Was AVS truly effective in enhancing Olympic security? Finally, what are the future prospects for AVS in France?
Article written by Louis Fourneau, student at 42 Paris, graduate of Sciences Po Paris
Don’t Miss It…
Press Review from 42AI’s Impact & Governance Hub.
OOPSY… — Want to hack Barack Obama’s Instagram account? Just ask Meta’s AI! (404 Media)
ILLEGAL ROBOCOP — When the French Police Illegally Use Facial Recognition
- Mass deployment of facial recognition on law enforcement officers’ phones (Disclose)
- EU Court of Justice rejects France’s fingerprinting database system (La Quadrature du Net)
COCORICO… Or Not? — Is France a champion of data centers or merely an “energy subcontractor” for the United States? Following investments announced Monday, June 1, at the Choose France summit, Le Monde raises questions (Le Monde)
PRESUMPTION… OF EXPLOITATION — A bill establishing a “presumption of exploitation of cultural content by AI providers” was adopted by the French Senate and sent to the National Assembly. “Concretely, under this law, creators, producers, or publishers will no longer need to prove their works were used; instead, AI players must prove the opposite.” (BPI France)
MEDIC-GPT — When AI Reasoning Matches Your Family Doctor’s (Science)
A GOOD FRIEND — Friendly AI Chatbots Make More Errors and Tell You What You Want to Hear (University of Oxford)
The Editorial Team Recommends: MATRIX: WINNING THE BATTLE FROM THE INSIDE
In the future, artificial intelligence has taken control over humans and uses them as an energy source. This synopsis could belong to a dozen works of fiction. But few have impacted science fiction as profoundly as The Matrix.
The trilogy (which became an anecdotal quadrilogue in 2019) directed by the Wachowski sisters depicts a dystopia where our daily lives are merely a virtual simulation created by computers to subdue humanity. One day, the computer programmer Neo learns the truth and joins the rebellion to regain power over his existence. At a time when artificial intelligence is on everyone’s lips and chatbots in everyone’s hands, it is difficult not to see an echo with our modern era.
The first installment, released in 1999, appeared as a cinematic UFO. The film’s syncretism stands out. First, for its masterful blend of Hong Kong martial arts, latex trench coats, and malevolent AI. Then, because it departs from typical representations of machines. In this post-apocalyptic world, the technology that surpasses man is not a robot but a fluid machine that does not physically resemble a living being.
The film not only warns against the pitfalls of misusing machines but goes further, prompting philosophical reflection. The Matrix depicts two worlds: one real, the other virtual yet indistinguishable from reality. Thus, confusion between these universes becomes possible. Those who indulge in the metaverse might well be tempted to never return to reality.
The Matrix also tells us that we have a choice. Between the “red pill” and the “blue pill,” Neo chooses red; he chooses to know the truth, even if painful. As a computer programmer, he will wage the battle from within the systems, refusing to leave decision-making power to the giants who shaped them. Food for thought.
Is the Wachowski sisters’ quadrilogy already in your top 10? No panic. The Matrix is a work we can constantly rediscover. Just as there are as many perceptions of reality as there are living beings, there are infinite interpretations of the film.
Ecology, transgender identity, spirituality, or even conspiracy theories — you decide to watch this saga through your own lens on the world!
Stay updated on the latest 42AI news and be the first to read the next newsletter! Follow us on LinkedIn and Medium.
Have remarks or suggestions? Do you wish to contribute to this newsletter, the Impact & Governance Hub, or simply exchange ideas with us? Feel free to contact us:
- Benjamin Karas’ Discord (co-head of the newsletter)
- newsletter@42ai.fr
This newsletter was written in French without AI assistance, and translated in English with the help of Lumo AI.
Editorial team: Benjamin Karas, Nora Todeschini, Louis Fourneau
