Shaping Ethical AI (English version)

Picture it. An ordinary morning in Paris. Cerise, a seasoned developer, settles in at her desk… until her AI assistant, “Ada”, starts proposing sophisticated architectural improvements. Suggestions so relevant that they sometimes go beyond her own level of expertise.

This scene, far from fictional, raises THE question that haunts our time: are we on the brink of truly autonomous AI?

I take you behind the scenes of this silent revolution. Through the story of Cerise and Ada, discover:

The real (and surprising) capabilities of today’s AI The technical obstacles still holding back its autonomy The technological breakthroughs that change the game How an unprecedented human-AI collaboration is taking shape

Between educational fiction and technical reality, this podcast separates fact from fiction in the debate on AI autonomy. Because understanding these stakes means preparing for a future where intelligence, human and artificial, is reinventing itself before our eyes.

“This is not a zero-sum game where one must dominate the other, but a complex dance where each partner enriches the other.”

Find the full transcript of this episode, along with all the sources and plenty more, at: AI at the threshold of independence: myth or reality?

Algorithms have neither morals nor intentions. But they learn from us. From our data. From our past decisions. And sometimes, without our wanting it, they inherit our deepest prejudices.

In this extract, I invite you to dive into the map of our digital wanderings: a journey through the invisible biases that silently sculpt our machines’ decisions… and already influence our lives. Recruitment, credit, justice, health: no field escapes.

Whether it is historical bias, representation bias or blind automation, every algorithmic distortion acts as a warped mirror of our society. It is not just a question of code: it is a matter of conscience.

And what if, to make our AI fairer, we first had to learn to look at ourselves more honestly?

Find the full transcript of this episode, along with all the sources and plenty more, at: Biases in AI are a disturbing reflection of our humanity

Some time ago, I spoke here about the danger of autophagy, the moment when artificial intelligence starts feeding on its own content, endlessly recycling the same ideas, impoverishing the diversity of knowledge.

But there is another risk, even more intimate: that of relieving us of the very desire to think.

Imagine an architect of knowledge. Every day, he sketches, questions, connects ideas. Then one day, a machine offers him the plans. Clear, fast, appealing. So he adjusts. He approves. But he no longer doubts.

AI does not attack us. It helps us. And that is precisely where the slide happens. It spares us the effort, and that effort is sometimes all we had left to remain truly human.

And what if the greatest danger lay not in the tool… but in the combination of these two phenomena?

An AI going round in circles on itself.

Humans who no longer wish to produce anything else.

Here I try to explore this question in an in-depth piece. Neither dystopian nor starry-eyed. Just clear-sighted.

Find the full transcript of this episode, along with all the sources and plenty more, at: When AI thinks for us, should we fear the disappearance of doubt?

Imagine an on-call doctor desperately searching for the medical history of an unconscious patient. Everything has been entered somewhere: in a surgery, a clinic, a specialist software package. But nothing flows. Nothing talks to anything else.

That is the true face of missing interoperability.

This seemingly technical word is in fact a major lever for transformation, at the crossroads of care, digital sovereignty, medical ethics and collective intelligence.

Here I explore why sharing medical data is not a luxury or progress for insiders, but a vital necessity. And above all, how we could act differently, at the level of carers, patients, institutions… and the technologies themselves.

Find the full transcript of this episode, along with all the sources and plenty more, at: When systems don’t talk, it’s the patients who suffer

Imagine a doctor whose every move is enriched and augmented by artificial intelligence, without ever losing the human warmth of a traditional consultation. That is the promise of “phygital”, this subtle fusion of the physical and the digital, which could now redefine medical practice.

But how can this alliance really transform your relationship with your doctor, improve diagnoses, and even tackle the challenge of medical deserts?

Discover how phygital does not replace people, but gives them back their full place in modern care.

A quiet revolution, where technology and tradition come together to offer medicine that is more personalised, more accessible, and deeply human.

Find the full transcript of this episode, along with all the sources and plenty more, at: Is PHYGITAL the future of healthcare?

Every day, without even noticing, we scroll, click, like… and feed our minds a lukewarm gruel of recycled information, mass-produced and drained of its marrow.

It is no longer only AI going round in circles in its own soup of synthetic content: it is us. Our brain, once curious, nimble and fond of complexity, now settles for digital crumbs, digested for us by machines.

The result?

An impoverished, standardised way of thinking, which is gradually losing its ability to tell true from false, the profound from the superficial.

This is not science fiction. This is about you. About me. About all of us. And about this invisible feast where humanity gorges itself on emptiness, one easy piece of content at a time. Ready to look squarely at what we are becoming?

Find the full transcript of this episode, along with all the sources and plenty more, at: Cognitive autophagy, when humans feed on impoverished content

With Trump’s return and his “America First” doctrine, our technological dependence is no longer a mere commercial disadvantage: it is a STRATEGIC TRAP closing in on us.

When Broadcom raised VMware prices by 300% (even 1,050% for some), it is not just an IT department problem: it is OUR MONEY filling the hole while our public services run on a shoestring.

The urgency to act has NEVER been so palpable.

Find the full transcript of this episode, along with all the sources and plenty more, at: 72% of our digital sovereignty under American flag: the urgency of a European cloud

While ChatGPT writes our emails, Midjourney illustrates our presentations and other AIs write our code, we find ourselves facing a fascinating paradox: never have we had such powerful tools to amplify our creativity, and yet never have we seemed so vulnerable to the consequences of that amplification.

In my article “Ethical challenges and future perspectives of generative AI”, I explore the grey area where neither starry-eyed techno-optimism nor paralysing catastrophism is enough to grasp reality.

I invite you on a journey to the heart of the four fundamental ethical dilemmas of these technologies:

But above all, beyond the diagnosis, I sketch out a “conscious digital pact” that could turn these challenges into a historic opportunity.

Because the question is not whether AI will replace us, but how to preserve our creative autonomy in a world where these tools amplify some of our abilities while risking the atrophy of others.

Find the full transcript of this episode, along with all the sources and plenty more, at: Ethical challenges and future perspectives of generative AI

Behind every ChatGPT query lies a reality few of us want to face: a colossal infrastructure devouring water, electricity and rare metals at a frantic pace.

I went behind the scenes of this technological revolution and the figures are dizzying:

A single ChatGPT query uses 5 TIMES more electricity than a standard Google search

By 2030, AI could generate 2.5 MILLION tonnes of electronic waste a year

A heavy ChatGPT user generates almost 300g of CO2 a year, multiplied by hundreds of millions of users…

But beyond the alarming figures, in this piece I explore concrete avenues: specialised models, distributed computing, frugal innovation… Because I am convinced of it: the future of AI will be green, or it will not be at all.

This extraordinary technology deserves better than our collective blindness. It deserves deep reflection on its sustainability.

Find the full transcript of this episode, along with all the sources and plenty more, at: Generative AI or the hidden face of a technological mirage

We love these tools. One prompt, and presto: a stunning image, a crystal-clear text, working code. A technological miracle, with no bill to pay.

But who is really paying for this algorithmic magic?

Behind every ChatGPT answer, every Midjourney visual, lie voracious data centres, colossal infrastructure and staggering energy costs. And yet we go on believing it is free… Will this illusion last? When will the bill arrive? And above all, who will fund this revolution?

This podcast is part of a three-part reflection exploring the little-known stakes of generative artificial intelligence, beyond its apparent free-of-charge nature.

➤ The illusion of a cost-free digital world

Generative AI, the hidden side of a technological mirage

Who should pay? Ethical stakes and future prospects

Find the full transcript of this episode, along with all the sources and plenty more, at: Generative AI or the illusory alchemy of digital free access