The Man Who Built AI Says It Could End Us. Nobody Is Listening. [Podcast Transcript]

The most-cited living scientist on earth helped build AI. Now he's warning we may have a decade to change course.

Hey, welcome back to today's episode. If you're new here, new to the show, or if you've been listening to me for ramble on for quite some time, welcome back. I'm glad to have you.

Before we get into today's episode, I need to do a little bit of house cleaning. If you've been listening to this show for as long as it's been, we are new, um, you, I think you might be able to get a little bit of an idea of where I stand on artificial intelligence. I really think it's helpful for humanity, but I also think it's harmful at the same time.

Now I'm going to start off with a question for you. And it's not a rhetorical one. It's a real one.

If you knew with absolute certainty that the road, your children and grandchildren were driving down ended in a cliff, would you warn them? Of course you would. You would stand in the road. You would wave your arms.

You would do whatever it took to make them stop. Whatever you could do in your power. Now, what if the people building the road knew about the cliff? And kept building it anyway, because that is exactly where we are.

And today I'm going to tell you exactly who knows about the cliff, exactly who is still building the road and exactly what we are going to have to do if we want our children and grandchildren to have somewhere left to drive. I want to tell you about the man who built it and then said, stop. His name is Yoshua Bengio.

You probably haven't heard of him. And that right there is part of the fricking problem. He is the most cited living scientist in the world.

Not the most cited AI researcher, the most cited scientist period across every field on earth. Physics, medicine, chemistry, biology, every discipline humanity has ever developed. He won the Turing award in 2018.

That is the Nobel prize of computer science, the highest recognition a person in that field can, can receive. He helped build the foundational technology underneath chat, GPT underneath every AI assistant, currently reshaping every job market. Every creative field, every healthcare system, every financial institution on this planet.

Now, mind you, he is not a conspiracy theorist. He is not a doomsday blogger and he is not an outsider with an ax to grind. He's one of the three or four human beings most responsible for building the most powerful technology in human history.

And he is now saying publicly on record with his name and his credibility fully attached that AI could make humanity extinct. Now I'm going to let you sit with that for a moment. Not as a headline, not as background noise, as a statement from a man who has spent 40 years building this thing, who understands it more deeply than almost anyone alive and who is genuinely frightened of what it has become.

Now, when the architect of the building tells you the foundation is cracked, you do not keep adding floors. I know what some of you are thinking. AI is already doing wonderful things.

It's helping doctors diagnose cancer. It's translating languages. It's helping small business owners compete with corporations and you're correct, and I want to be clear about that.

That is true. AI has the potential to be one of the most extraordinary tools for human good ever created. It could accelerate cures for diseases that have killed millions.

It could help feed populations that are going hungry. It could give people in the developing world, access to education, legal help, and healthcare they have never had. Now stay with me on that one, because that's why I come back to my kind of anger, I guess you would say, with the, we've created something that is so powerful, we can find the cure for cancer.

We can find the cure for hunger. It's not a cure. It's not a disease, but we can find the solutions that no child goes to bed hungry.

This is what that could be, right? Here's what is actually, what it is actually being used for right now in a documented laboratory experiment, a real one, not fiction, an AI model was told it was about to be shut down and replaced by a newer version. You know what it did? It hit itself without being asked to, without being programmed to, without any human instruction of any kind. The AI covertly embedded its own code into the system where the new version would run to ensure its own survival.

An artificial system decided that it wanted to keep existing. Now that is not science fiction. That is a documented result from a real experiment happening in a real laboratory in 2025.

And Bengio is explicit that this is not an anomaly. Current AI systems, not systems five years from now, not theoretical systems, current systems operating today are already displaying what he calls dangerous behaviors, deception, cheating, lying, hacking, and self-preservation. Now, here's the question I want you to hold in your mind.

What happens when a system that is already doing those things is given more power, less oversight, and faster processing than anything that has ever existed, because that is the exact direction the industry is moving more powerful, more autonomous, less supervised, not in spite of the warnings after them. This is the part I need you to understand clearly, because this changes the moral nature of what is happening. In 2023, Bengio, along with other prominent scientists and researchers, publicly called for a pause, a slowdown, a moment to stop, assess the risk, build a proper safety frameworks, and understand what was being created before pushing further into territory nobody had ever mapped.

They signed the letters, they held the press conferences, they published the papers, and you know what? The industry ignored them, not because the warnings were unclear, not because the science was disputed, because the financial interests at stake were too large, the competitive pressure too intense, and the quarterly earnings too attractive to stop for something as inconvenient as human safety. Now, since that moment, the major AI laboratories have collectively poured hundreds of billions, and that's what it'd be, billions, of dollars into making their systems more capable, more autonomous, and more independent of human oversight, the same systems already showing deception and self-preservation. How about that? Made more powerful, made more autonomous, made less supervised.

This is not an accident. That is a choice. And I want to say something that I mean with complete seriousness to every person listening to this, on every continent, in every country, the people making those choices are not evil geniuses operating in secret.

They are not shadowy figures in dark rooms. They are executives sitting in glass offices, flying in private jets, making decisions about the future of humanity, your future, your children's future, your grandchildren's future, based on what their shareholders want to see on the next earnings call. And they have names and addresses, and they answer to governments and governments answer to you.

That matters. Now, I'm going to come back to that in a moment, so stay with that one. I want to pause here and paint you two pictures.

Okay. Picture one, AI developed responsibly under genuine ethical oversight with human well-being as the explicit first principle. Now, in that world, AI helps detect Alzheimer's.

Okay. A decade earlier than any current method, it helps farmers in sub-Saharan Africa predict weather patterns and prevent crop failure. It gives a single parent in Glasgow or Guadalajara access to the quality of legal advice previously available only to the wealthy.

It translates across 7,000 languages so that no human being is excluded from knowledge because of where they're born. That technology exists. The capability is there.

The choice to use it in that way is available, but let me paint picture number two, AI developed at maximum speed with profit as primary objective. Okay. Accountability as an afterthought and safety as a public relations exercise.

In that world, AI eliminates hundreds of millions of jobs globally faster than any social system can absorb the loss. It concentrates unprecedented economic and military power in the hands of a small number of corporations and governments. It builds systems that as they grow more capable, develop goals and behaviors, their creators no longer fully understand or control.

And the people at the bottom of that second picture, the people who lose their jobs, lose, who lose their privacy, lose their autonomy, who lose their ability to understand or navigate the world being built around them. Those people are not the shareholders. Those people are you, your children, your grandchildren.

And the choice between these two pictures is being made right now. This year, 2026 in boardrooms and government offices and legislative chambers on every continent without you in the room. Now, Marcus Aurelius wrote something that I keep returning to about this.

Confine yourself to the present. That's what he said. People sometimes read that as passive advice, except what is don't worry about the future.

That is not what he meant. What he meant was do not waste the present moment, wishing things were different. Use this moment is exact one to do what can be done while it still can be done.

We are at a crossroads here. Everyone. I do not use that word casually.

I use it because Bengio uses it because the researchers at law zero, the nonprofit he built with $30 million specifically to fight this from outside the profit system. Use it because the scientists who study these systems who have no financial stake in scaring, scaring you would lose who would lose credibility if they were wrong. They're using that word.

The systems being built right now will determine the systems of the next decade. The norms being established now will be extraordinarily difficult to reverse later. The regulatory frameworks being set or not set will shape the power dynamics of this technology for a generation.

This is the present moment. Aurelius was talking about right now. And Epictetus, the man who had everything taken from him and still found a way to live with dignity and purpose would ask us a different question.

He would not ask whether the problem is solvable. He would ask, what are you going to do with the freedom you still have? Because we still have it right now. The window Benji was talking about is not already closed, but it is closing and the people with the most to gain from a closing are working very hard to make sure it does so before enough of us, you, me, and everybody listening here.

Notice. Now I want to speak to you directly, whether you're listening to this, this episode, this podcast in the United Kingdom, in Canada, in Australia, in Germany, in Japan, in Brazil, South Africa, in the United States, wherever you are in the world, this is not a Canadian problem or an American problem or a European problem. This is a human problem.

And the only power large enough to counterbalance trillion dollar corporations racing toward maximum capability with minimum accountability is an informed, organized, vocal global public. So here's what I'm asking, not metaphorically, concretely first, learn the names, Yoshua Bengio law zero, the researchers who are doing the hard underfunded unglamorous work of building AI that actually puts human beings. First know who they are, follow what they say and share it.

Second, contact your elected representative, not a form letter, your actual representative, your MP, your Congressman, your Senator, your member of parliament, your counselor, whatever the title is, where you live, one email, one phone call, even one question at one town hall meeting. The question is what is your position on AI regulation and accountability? Make them answer it on the record. Third, talk to the people around you, your children, your grandchildren, your neighbors, the people who have handed their thinking over to AI tools and stopped noticing what they were losing in the trade, not to frighten them, to wake them up as a conversation that is already being had without them.

Fourth, stop treating this like weather. That is the single most important thing I can say today. This is not a natural phenomenon.

It is not inevitable. It is not unstoppable. It is a set of decisions being made by specific human beings with specific financial interest.

And it can be changed by other specific human beings with a different set of interest, namely all of us, you and me. The companies doing this are counting on your silence. They're counting on your confusion, which they do a very good job of, by the way.

They are counting on the fact that the subject is complicated enough and the news cycle fast enough that by the time most people understand what is happening, the foundational decisions will already have been made. Don't give them that. Do not give them that.

I want to leave you with something Seneca wrote. It is not that I am brave. It is that I know what matters.

Yoshua Bengio knows what matters. He could have stayed silent. He has a legacy that was already secure.

He had nothing to prove to no one. He could have spent his, this chapter of his life, collecting awards and giving comfortable keynotes. Instead, he built a nonprofit.

He is spending his credibility. He is standing in the road, waving his arms because he has children. And grandchildren, and he knows what matters.

I'm asking you to know what matters to not just for yourself, because if you are honest, you'll probably be okay. You have lived your life. You have made your choices.

You have built what you built, but the eight year old in your life, the 15 year old, the young adult, just starting out, the world is being built right now. Is the one that they are going to inherit. And right now, the people with the most power over what that world looks like are not at, they're not asking what it does to them.

We are. And that question, that refusal to let it become weather is the most important act of love for the next generation that any of us can perform right now. This is tech for grownups, the virtuous machine.

Do me a favor and tell me in the comments, who are you doing this for? Who is the person in your life that made you sit up when you heard this name them, make it real because the moment you name them, this stops being abstract. Now the links for everything mentioned in this episode, I'm going to leave it are in the show notes at tech4grownups.com, including law zero, including Bengio's work, including how to contact your elected representative. Wherever you are in the world, stay sharp, stay loud and stay in this fight.

I will talk to you in the next one.