kipode.de Transatlantik
Transatlantic · The Diary of a CEO

Brake Before It Is Too Late: Bernie Sanders on AI

Bernie Sanders has sat in the US Congress for 35 years and twice fought for the Democratic presidential nomination. With Steven Bartlett he explains why he considers AI the most consequential technology in human history, why he wants to ban superintelligence and why the public should own half of the big AI companies.

November 19, 2026 Bernie Sanders · Steven Bartlett 1 hr 30 min original length · condensed AI section

The video only loads from YouTube after you click it. No connection is made beforehand.

On 19 November 2026, Steven Bartlett, the British entrepreneur and host of The Diary of a CEO, released an hour-and-a-half conversation with Bernie Sanders. The senator from Vermont talks about the midterm elections, big money in politics, health insurance, taxes and the war against Iran. Around a quarter of an hour, roughly a seventh of the episode, is about artificial intelligence. We cover only that part here, and we have checked his figures. Watch from 0:34:47

The guestWho is speaking

Sanders is 85 and has sat in Congress since 1991, first in the House of Representatives and since 2007 in the Senate. He belongs to no party but usually votes with the Democrats, and he ran for their presidential nomination in 2016 and 2020. His life’s theme is inequality of income and wealth.

Sanders says himself that he is no tech expert, quite the opposite. But he reads a lot, and it became clear to him that he had to take up the subject. Watch from 0:35:55 Keep in mind while reading: Sanders is a politician, and in 2026 he introduced two bills of his own on AI, which he promotes in the conversation.

The machinesA robot needs no minimum wage

Bartlett steers the conversation from minimum wage and workers’ rights to the machines. Many minimum-wage workers, he says, are in exactly the jobs most exposed to AI and robots. A robot has no union, needs no minimum wage and no holiday, and works around the clock. Watch from 0:34:47 He describes a video in which a robot worked next to a human on a production line for seven days and outperformed him. Watch from 0:35:28

Sanders answers in principle. AI and robotics, he says, are the most consequential technologies in human history; next to them the industrial age looks like a slow process. Watch from 0:36:12 He waited to see what his colleagues on the relevant committees would do about it, and the answer was nothing. So he got involved himself.

The workHow many jobs will be left?

First, Sanders concedes that nobody knows how things will turn out, and that smart, honest people disagree. Watch from 0:36:40 But some heads of large AI companies, he names Anthropic, talk about many millions of jobs that could disappear. Watch from 0:36:53 Jeff Bezos, the founder of Amazon, is putting together a fund of 100 billion dollars to automate factories. That means robots taking over work that people used to do. Watch from 0:37:12

In Texas, heavy trucks are already driving on the roads without a driver. Transport, he says, accounts for 67 million jobs, taxi, bus and ride-hailing drivers. How many of them will be left in ten years, nobody knows. Watch from 0:37:27 And young graduates are already struggling to find a first job. Watch from 0:37:44

The driversWho actually wants this?

Then Sanders asks the question that matters most to him: who is pushing this? Nobody on the street wants a technology that wipes out their job. It is being pushed by the richest men in the world, he says, naming Elon Musk, Bezos and Larry Ellison, the founder of the software company Oracle. Watch from 0:38:00 Bartlett objects that their statements often contradict each other over the years. Sanders sees it more simply: they want more wealth and more power. Watch from 0:38:37

Musk does now and then promise a kind of paradise: work will no longer be necessary and everyone will get a cheque from the government. Where the money is supposed to come from if nobody works anymore remains open. Watch from 0:38:48

FreedomSurveillance, loneliness, war at the push of a button

For Sanders it is not only about jobs. He mentions Flock, an American company whose cameras record car number plates and analyse them for police departments. Watch from 0:39:02 He thinks it likely that soon every phone call, every message and every website visited will be recorded somewhere. Bartlett agrees. Watch from 0:39:14 What does that mean for freedom, Sanders asks, when someone knows where you drive and whom you meet?

Then the question of what happens to people themselves when children spend a lot of time with conversation partners that are programs and tell them about their heartbreak. Watch from 0:39:43 In his office in Vermont, young staff take calls from citizens, and many find it hard even to talk on the phone because they only send messages. Watch from 0:40:03 Bartlett adds that fewer people are entering relationships and mental health problems are rising.

War is changing too. Soldiers used to fight; in future someone presses a button. If a president no longer has to fear young soldiers coming home dead, what does that do to his readiness for war? Watch from 0:40:39 Finally, Sanders refers to Geoffrey Hinton, one of the founders of modern AI and winner of the 2024 Nobel Prize in Physics. Hinton considers it quite possible that AI becomes smarter than humans and can no longer be controlled. Watch from 0:41:05

The banNo machine smarter than us

Bartlett asks about Sanders’ demand to ban the development of superintelligence, meaning an AI superior to humans in everything, and to pause advanced AI for now. Is that a permanent ban, and what does it mean in the race with other countries? Watch from 0:41:27

Sanders knows his colleagues’ objection: “Bernie, you just want China to win.” Watch from 0:41:54 He has spoken with Chinese scientists, and some share the same fears. A rational government would call President Xi and negotiate an agreement. He recalls Ronald Reagan and Mikhail Gorbachev, who concluded nuclear arms treaties despite opposing world views. Watch from 0:42:17

His own line: he does not want to develop a technology that is smarter than humans. For that, a ban, and for advanced AI a pause until there are sufficient safeguards for people and workers. Watch from 0:42:40 Anyone who can show that an application does good should go ahead. But anyone who wants to wipe out 20 million jobs should not. Watch from 0:43:07

Bartlett plays devil’s advocate: if the US slows down but China, Russia or Europe do not, won’t America lose even more jobs, as if it had missed out on the internet? Watch from 0:43:31 Sanders concedes that this is not an easy question. With China in particular, there needs to be an agreement that does not just sound good but whose compliance can actually be verified. Watch from 0:44:01

OwnershipHalf for the public

Bartlett raises Sanders’ second demand: the state should own 50 percent of the big AI companies. Watch from 0:44:16 Sanders justifies it with the same idea. He does not want Musk, Bezos or Sam Altman, the head of OpenAI, to make decisions on their own that change the lives of tens of millions of people. If representatives of the public sat on half of the boards, that would slow development down and the companies would be held accountable. Watch from 0:44:42

The entrepreneurWhat Sanders values in Musk and what he does not

Later, Bartlett asks whether Sanders respects Elon Musk. Sanders praises his hard work and intelligence but criticises his power: Musk owns the platform X, has enormous influence in robotics and AI, and pours a lot of money into elections. Watch from 1:09:33

Sanders does not see Musk’s promise that nobody will have to work anymore because AI does everything as a good prospect. He asks Bartlett whether he enjoys his work. Bartlett says yes. Then surely he would not want a robot to take it over just because it can, while he tends his garden. Watch from 1:10:46 If someone invents a cure for cancer or Alzheimer’s, wonderful. But innovation for the sake of innovation is not automatically good. Watch from 1:11:16

Bartlett says he has interviewed many AI experts. They do not agree about the future, but when he reads their statements from the last ten or fifteen years, almost all of them point to an uncomfortably high probability of a bad outcome. Watch from 1:11:40 Sanders: then take a deep breath and ask what on earth we are doing. Watch from 1:11:50

The last questionWhat if the data centres stood still?

At the end of each episode, Bartlett’s guest leaves a question for the next one. This time it comes from a guest from the world of finance who knew Sanders was next: is he aware that the US economy would fall into an immediate recession if all AI data centres were switched off tomorrow? Watch from 1:28:29 Sanders says he is aware of that. But investors should have thought about it before pouring huge sums into a technology whose own company heads say they do not know where it is going and that it could be very dangerous. Watch from 1:28:40

The fact checkWhat can be verified

Jeff Bezos’s fund so far exists only as a report. According to the Wall Street Journal, Bezos held early talks in March 2026 about raising 100 billion dollars to buy industrial companies and automate them faster with AI. According to Axios, the fund is linked to Project Prometheus, an AI company Bezos co-leads. People close to it stress that it is about development and prototyping, not about replacing workers.

Driverless trucks really do exist in Texas. The company Aurora has been hauling freight regularly between Dallas and Houston since the end of April 2025 with nobody in the cab, according to TechCrunch. The 67 million transport jobs, on the other hand, cannot be confirmed. According to the US Department of Transportation, around 11 million people worked in transport occupations in 2024. Sanders gives no source in the conversation for the 20 million jobs.

Sanders has indeed introduced both demands as bills. On 18 June 2026 he presented the American AI Sovereign Wealth Fund Act: AI companies with more than 200 million dollars in annual AI revenue would hand over half of their shares to a public fund once. In the conversation Sanders mainly talks about board seats; the bill is about voting shares. He introduced the Ban Artificial Superintelligence Act on 23 September 2026 together with Representative Greg Casar. How it came about is explained in our AI Radar from September (in German). Neither bill has been passed.

ContextWhat This Means for Europe

Sanders’ idea of putting public representatives on the boards of large corporations sounds radical in the US. Germany has had something related for almost 50 years. Under the Codetermination Act of 1976, employees fill half the supervisory board seats in companies with more than 2,000 staff. In a tie, however, the chair decides, and the chair is usually chosen by the shareholders. So it is employees who sit at the table, not the state, and the final say stays with the owners.

On the surveillance Sanders fears, the EU has already drawn lines. Since February 2025, Article 5 of the EU AI Act has banned, among other things, scoring citizens by their social behaviour and, with narrowly limited exceptions, real-time facial recognition by police in public places. European law, however, has no ban on superintelligence of the kind Sanders calls for. The regulation governs how AI is used, not how powerful it may become.

For European readers, one question remains open that Sanders asks himself: can a single country or a union of states stop development if others carry on? His answer is a verifiable agreement with China. How AI safety experts see the same question is shown in our episode with Daniel Kokotajlo.

Summary and context by kipode.de. The text is our own; statements from the original are timestamped. The video is from The Diary of a CEO and is embedded via the official YouTube player. Not an official translation, no connection to the podcast.

About Transatlantik

What is said over there, readable over here.

Transatlantik is the kipode.de desk for American debates about artificial intelligence. We listen to the conversations in full, summarise them in our own words and add timestamps into the original, so every statement can be checked.

At the end comes the question nobody over there asks: what does this mean for us in Europe? Every episode is available in German, English, French and Spanish.

4languages per episode
100 %our own text, not a translated transcript
0downloaded videos or images — only the official player

See all episodes →