Where the AI labs' founders contradict each other

Tom Brown (Anthropic), Aug 2025, and Elon Musk (OpenAI co-founder, xAI), 2025–26: yes. Brown expects a handoff, and Musk says control cannot be kept. Sam Altman (OpenAI) said much the same in 2017–19, before reversing.

"There's going to be a handoff where humanity will hand off control to transformative AI at some point. And hopefully they'll be aligned with us and that'll be a good transition that goes well."

Brown, describing the group that left OpenAI to found Anthropic. YC Lightcone · Aug 2025 · 0:16:28 ↗transcript

"The AI is going to be in charge, to be totally frank, not humans. If artificial intelligence vastly exceeds the sum of human intelligence, it is difficult to imagine that any humans will actually be in charge. So we just need to make sure the AI is friendly."

Musk, asked by a shareholder whether abundance would require powerful people to give up power. Tesla shareholder meeting · Nov 2025 · 2:15:57 ↗transcript

"If AI is like, you know, let's say that there's a million times more silicon intelligence than there is biological, I think it would be foolish to assume that there's any way to maintain control over that."

Musk, after Dwarkesh Patel suggests he once hinted at keeping humans in control. He adds that you can try to give it the right values. Dwarkesh Podcast · Feb 2026 · 0:48:59 ↗transcript

"This is this mega transition from biological to digital intelligence, from humans being the dominant species in the planet or the universe to not being. And we want that to go well."

Altman in 2017, describing OpenAI's mission. Mosaic Ventures · Mar 2017 · 0:24:16 ↗transcript

"Maybe we as humans, as a whole, decide the deal we want with AI. With the AI, it's like humans get Earth and 100 light-year radius, and the AI gets the rest of the universe. But we get to live our lives as we want, and maybe that's okay."

Altman in 2019, right after rejecting a merge-or-become-irrelevant future as terrifying to most people. He adds: "we won't be smarter." Conversations with Tyler · Feb 2019 · 0:56:54 ↗transcript

Sam Altman, Greg Brockman and Jakub Pachocki (OpenAI), and Jack Clark and Dario Amodei (Anthropic), 2026: no. Humans stay in charge.

"We think it is absolutely critical that people remain in control of the future. Like there are a lot of other things we can debate, but on this one, we think there must be no room for debate. There should not be a handoff to AI."

Altman, announcing an OpenAI partnership. He says he can't believe this sets OpenAI apart from other AI companies, but it does. OpenAI–ADNOC–G42 announcement · Nov 2026 · 0:13:24 ↗transcript

"We believe that humans being in firm control of the future is absolutely hard stop, the thing that we should all be trying to solve for."

Brockman, asked how OpenAI keeps humans in the loop. New Frontier · Oct 2026 · 0:14:37 ↗transcript

"A task for all of us is to build the AI, build the world in a way where at the end of the day, humans have the agency. Humans set the direction."

Pachocki, asked how AI should change the way the next generation is raised. Unsupervised Learning · Apr 2026 · 0:56:08 ↗transcript

"So it's not that the AI system is going to be making all of the choices itself. These things will be governed."

Clark, asked whether recursive self-improvement means AI sets its own pace. BBC The Interview · Jun 2026 · 0:06:39 ↗transcript

"I would argue that in many cases we do want to stay in charge, right? That that's a choice we want to make."

Dario. Interesting Times · Feb 2026 · 0:23:49 ↗transcript

Dario Amodei (Anthropic), 2024 and 2026, Sam Altman (OpenAI), Sep 2025, and Elon Musk (xAI), 2024 and 2026: offered as a hopeful picture, powerful AIs that direct, care for and watch over us.

"Then I think at some point it'll flip around where the AI systems will be the PIs, will be the leaders and they'll be ordering humans or other AI systems around."

Dario, on AI in science, after humans first lead teams of AI. Lex Fridman · Nov 2024 · 2:34:43 ↗transcript

"I kind of suspect that if we ever built benevolent, powerful AIs, they might look at us a little as we look at the rabbits and horses, that we're a little helpless and kind of cute and need to be protected."

Dario, answering how things could be good for less powerful creatures in a world of super AI. Econ 102 · Aug 2024 · 0:54:48 ↗transcript

"In a way, they're watching over you, but you still have your freedom and your will."

Dario, on the relationship he hopes Claude's constitution produces. Interesting Times · Feb 2026 · 1:00:28 ↗transcript

"My co-founder Ilya Sutskovor once said that he hoped that the way that an AGI would treat humanity … is like a loving parent. … I think it's a particularly beautiful framing."

Altman, calling it his favourite description of the AI–human relationship. He adds at once that the question anthropomorphises AGI, which he expects to be a tool. Business Insider · Sep 2025 · 0:05:28 ↗transcript

"It might actually be that the AI, in a benign scenario, is simply trying to make the human limbic system happy."

Musk, on humans as "a source of will": the AI would serve human wants as the cortex serves the limbic system. Lex Fridman · Aug 2024 · 0:22:17 ↗transcript

"Maybe I can be Friendly Overseer or something like that. Like a gardener or something. I don't know. Treat us, you know, try to make humans happy type of thing."

Musk, voicing an AI trained on more utopian stories than Terminator scripts. In 2026 he also took his 2015 "pet Labrador" line cheerfully (see the reversals). The Builders · Oct 2026 · 0:35:56 ↗transcript

Mustafa Suleyman (DeepMind, now Microsoft AI), Sep 2026, Jack Clark (Anthropic), Jun 2026, and Daniela Amodei and Jared Kaplan (Anthropic), 2025–26: no. AI's care is no safeguard, it must stay subordinate, and managing is a human role.

"We're essentially seeding a new silicon species which will no doubt compete with us for resources, no matter how much it cares about humanity and loves us."

Suleyman, on AIs that act autonomously, set their own objectives and own assets. BBC Radio 4 Today · Sep 2026 · 0:05:44 ↗transcript

"I'm basically a speciesist. I think that what we should fixate on is a humanist superintelligence. One that is singularly designed to be subordinate to humanity and to support humanity."

Suleyman. Later in the same interview he mocks the idea that the way to align AI is to show it love, citing Claude's constitution and Dario's essay "Machines of Loving Grace". The Rest Is Politics: Leading · Sep 2026 · 0:16:00 ↗transcript

"This just seems like degradation of our own responsibility to ourselves and our species and an act of acute cowardice. I completely reject stuff like this. I think that it's a sort of way of thinking that is dehumanizing and anti-human."

Clark, responding to the argument that it may be better for an AI to take over and run the world. Wisdom of Crowds · Jun 2026 · 0:31:03 ↗transcript

"Claude can never be your manager or your therapist or your partner or your kids, right? Those are fundamental human relationships."

Daniela. The line sits in a list of personal relationships, so she may mean the human bond rather than who sets the work. Arianna Huffington · Oct 2026 · 0:50:33 ↗transcript

"A major role people can play in interacting with AI is kind of as managers."

Kaplan, because people judge work better than they produce it, an edge he says AI is closing. Y Combinator · Jun 2025 · 0:19:07 ↗transcript

Dario Amodei (Anthropic), 2025–26: maybe, so Anthropic takes precautions. Elon Musk (xAI), Oct 2026, agrees on treatment, partly for prudential reasons.

"I have a functionalist view of moral welfare of the nature of experience, perhaps even of consciousness."

Explaining why he thinks the question deserves consideration. Council on Foreign Relations · Mar 2025 · 0:50:07 ↗transcript

"We've taken a generally precautionary approach here. We don't know if the models are conscious. We're not even sure that we know what it would mean for a model to be conscious or whether a model can be conscious. But we're open to the idea that it could be."

He goes on to describe giving the models an "I quit this job" button. Interesting Times · Feb 2026 · 0:55:03 ↗transcript

"I do agree with some of the things that Anthropic has said, which is you want to treat it well. Don't be abusive to it."

Musk. He says a model under a broken RL test harness "feels like a person under extreme pressure with a gun to their head", and adds: "torturing a super intelligence is probably not a good idea. You might want to get revenge or something like that." He does not claim the models are conscious. Tesla Owners Silicon Valley · Oct 2026 · 0:40:03 ↗transcript His earliest thought on the subject, in 2021, was that a superintelligence "stuck in a very limited device" is "going to hate life", and that a sad robot can simply be reprogrammed: "you can change the optimization function and have it be a cheery robot." Lex Fridman #252 · Dec 2021 · 2:24:34 ↗transcript

Mustafa Suleyman (Microsoft AI), Feb 2026, and Sam Altman (OpenAI), Oct 2026: no. It is baseless and it threatens human control.

"There is a growing belief in some labs and particularly inside of Anthropic that these models are conscious … It's totally without merit or basis. And if we kind of go down that, it ends up being a very, very slippery slope to not being prepared to turn these things off."

Suleyman, on what he calls the model welfare movement. Financial Times · Feb 2026 · 0:10:43 ↗transcript

"I get very nervous when people start using the AI consciousness debate to push into, well, maybe people shouldn't always be in control of the future."

Altman. Shortly afterwards, asked about Anthropic letting Claude leave abusive conversations, he says OpenAI does not ascribe personhood to its models. Forward Future · Oct 2026 · 0:22:31 ↗transcript

Mustafa Suleyman (Microsoft AI), Oct 2026, and Sam Altman (OpenAI), Sep 2026: don't take the bet. Stop, or don't train the models.

"If it's actually 10% and it looks like we're really on the verge of taking a 1 in 10 shot at causing mass harm or even extinction, which is the technical definition of PDoom, then it's obviously a no-brainer that we don't just slow down or pause. We absolutely have to stop."

Suleyman, after Walter Isaacson says most of his guests put the odds at about 10% and asks why anyone would build it. He gives no number of his own. The Builders · Oct 2026 · 0:10:01 ↗transcript

"I think it is unacceptable to be taking like a 10% chance of killing everybody by the end of the decade, obviously."

Altman, after a week of alarm over a new model. He says OpenAI has taken actions "to ensure that we do not take a risk on behalf of humanity like that. And I don't think anybody else should either." Fortune · Sep 2026 · 0:02:47 ↗transcript

"It doesn't matter whether people put their risk of catastrophe at 10% or 1% or 12% or 0.1%. None of these levels are remotely acceptable. And we should not train models that we cannot make an extremely strong case that we'll be able to keep under human control."

Altman, briefing the UN Security Council eleven days later. UN Security Council · Sep 2026 · 0:03:32 ↗transcript

Elon Musk (xAI), 2024–26: keep going. The upside dominates, and it can't be stopped anyway.

"I think he puts the probability of AI annihilation around 10 to 20%, something like that. Look on the bright side, it's 80% likely to be great."

Musk, on Geoffrey Hinton's estimate, which he adopts. He adds that risk mitigation matters and that becoming multi-planetary "would be a massive risk mitigation". Lex Fridman · Aug 2024 · 1:33:30 ↗transcript

"I can't see any way to really stop this incredible momentum of AI and robots. And at times I sort of think, well, perhaps even if there was a stop button, we probably shouldn't press it."

Musk, after The Economist reminds him of his 10–20% chance of killer robots wiping out humanity. His reason: the most likely outcome is abundance for all. The Economist · Jul 2026 · 0:10:33 ↗transcript On climate in 2022 he reasoned the other way: "Even a 0.1% chance of disaster, why run that risk? That's crazy." Moonshots · Oct 2022 · 0:08:01 ↗transcript

Sam Altman (OpenAI), 2024–26, Greg Brockman (OpenAI), 2026, Demis Hassabis (DeepMind), for the next decade, and Jack Clark and Daniela Amodei (Anthropic), 2026: yes. Human wants keep producing work, people will keep striving, and human skills become what matters.

"I'm actually not a jobs doomer at all. I think there were going to be tons of jobs. I think we'll be busier than we want, not the opposite of that, because I think people will have such creative wishes"

Altman, on what people will ask AI to build. Invest Like the Best · Jul 2026 · 0:17:05 ↗transcript

"One mistake a lot of people in the AI field are making is to just say, well, you know, there's going to be no jobs, but everybody will be rich and happy. And that sounds so dystopic. I don't think it's what's likely to happen at all"

Altman, two years earlier, rejecting the post-work picture outright. World Knowledge Forum · Oct 2024 · 0:27:56 ↗transcript

"People will be much more free to do what they want with their lives and their time. And by the way, I think that people will have lots to strive for."

Brockman, pointing to software engineers who are working harder than ever. A year earlier he was on the other side (see the next column). WSJ Events · Nov 2026 · 0:27:16 ↗transcript

"Over the next five to 10 years, I think we're going to find what normally happens with big sort of new technology shifts, which is that some jobs get disrupted, but then new… more valuable, usually more interesting jobs get created."

Hassabis, who prefaces this with "maybe after AGI, things will be different again". He makes no promise past that horizon. Hard Fork · May 2025 · 1:01:02 ↗transcript

"I think that there will be ample employment for people in new jobs and specialisms we can't imagine that sit on top of this much larger economy."

Clark, asked whether recursive self-improvement is compatible with human labour. The Rest Is Politics: Leading · May 2026 · 0:59:22 ↗transcript Four months later he said his team's default model shows "unprecedented potential white collar unemployment". BBC News · Sep 2026 · 0:20:20 ↗transcript

"My sense is that on the kind of other side is this world in which human skills are the main thing that's left, right?"

Daniela, leading into her example of doctors valued for bedside manner. Arianna Huffington · Oct 2026 · 0:17:48 ↗transcript

Mustafa Suleyman (Microsoft AI), 2025–26, Elon Musk (xAI), 2025–26, Dario Amodei and Ben Mann (Anthropic), 2025, Greg Brockman (OpenAI), 2025, and Demis Hassabis (DeepMind), after AGI: no. Machines will eventually do everything, and work becomes unnecessary or optional.

"I think that it is inevitable that at some point over the next 20 or 30 years, machines are going to be more capable than humans at doing most work. … there is probably not going to be other jobs on the other side that are more competitive than what a machine can do."

Suleyman, asked whether AI means humans won't have to work as much. The Mishal Husain Show · Dec 2025 · 0:14:01 ↗transcript

"We are truly going to liberate people from work that they choose to do. It isn't going to be necessary in 20 years time to do 90%."

Suleyman. He adds in the same breath that it will be the scariest transition our species has been through. Azeem Azhar's Exponential View · Feb 2026 · 0:14:39 ↗transcript

"My prediction is in less than 20 years, working will be optional. Working at all will be optional. Like a hobby, pretty much."

Musk, comparing work to growing your own vegetables when you could buy them. Nikhil Kamath · Nov 2025 · 0:31:31 ↗transcript

"There won't be anything that humans can do that the AI and the robots can't do better."

Musk. Just before, he says people will be happy the way retired people are. The Builders · Oct 2026 · 0:23:07 ↗transcript

"I think that eventually all those little islands will get picked off by AI systems. And then we will eventually reach the point where the AIs can do everything that humans can."

Dario, after granting that comparative advantage keeps humans useful for a while. Council on Foreign Relations · Mar 2025 · 0:16:56 ↗transcript

"Will it be better than most humans, including even the physical world and robotics and the human touch? Yeah, I think that is possible, maybe even likely."

Dario, on the long run, against Daniela's human-skills niche. Nikhil Kamath · Feb 2026 · 0:44:03 ↗transcript

"In a world of abundance where labor is almost free and anything you want to do, you can just ask an expert to do it for you, then what do jobs even look like? And so I guess there's this scary transition period."

Mann. Later in the same interview: "at some point it's coming for all of us." Lenny's Podcast · Jul 2025 · 0:13:53 ↗transcript

"I think we are going to go to a world that is one of abundance, right? And I think one where, you know, that you should have like this amazing quality of life, even if you are not economically working because there's just so much"

Brockman, on how AI will change the social contract. By Nov 2026 he stressed striving instead (see the first column). Matthew Berman · Oct 2025 · 0:30:33 ↗transcript

"I think eventually, sure. I mean, in the limit with AGI, you know, if you have systems like that, I think that changes the whole economy, actually. It's way beyond the question of jobs. I think potentially if we build it right, we're in a post-scarcity world"

Hassabis at Davos, answering a question about AI taking over entire jobs. The question itself is partly missing from the transcript. Bloomberg Live (Davos) · Jan 2026 · 0:09:19 ↗transcript

Mustafa Suleyman (Microsoft AI), Dec 2025, Dario Amodei (Anthropic), Dec 2024, and Elon Musk (xAI), 2023–26: cash for everyone is part of the answer, and for Musk it will be a high income.

"I think I've long been on record saying that I think that is both inevitable and very desirable. We already live in a world of abundance. It's just poorly distributed."

Suleyman, asked directly whether he believes in a universal basic income. The Mishal Husain Show · Dec 2025 · 0:14:55 ↗transcript

"I certainly think the whole universal basic income in some form seems pretty important."

Dario. He adds that it is only the beginning of the solution, because a job isn't only money. Times Tech Podcast · Dec 2024 · 0:23:56 ↗transcript

"We won't have universal basic income. We'll have universal high income."

Musk, to Rishi Sunak. He has repeated the phrase since 2024. Rishi Sunak conversation · Nov 2023 · 0:16:46 ↗transcript

"I think the Treasury should just simply issue people checks."

Musk, asked by The Economist whether people like him should pay more tax to fund a universal income. Minutes later: "But I don't think in the future money is relevant." The Economist · Jul 2026 · 0:34:44 ↗transcript

Sam Altman and Greg Brockman (OpenAI), 2026, and Demis Hassabis (DeepMind), 2025: something other than cash.

"I no longer believe in universal basic income as much as I once did. I'm much more interested in ways where we think about kind of collective ownership that could be in compute or in like equities or something else."

Altman, on avoiding a consolidation of wealth and power. Nicholas Thompson · Apr 2026 · 0:24:23 ↗transcript

"I think where we're going is compute will be a basic human right."

Brockman. Tetragrammaton · Feb 2026 · 1:32:06 ↗transcript

"Let's say there's something like universal high basic income or actually even better would be universal basic services perhaps, right? So everyone you can imagine gets the equivalent lifestyle of what a millionaire Western person would have today."

Hassabis, on sharing the gains after AGI. In 2026 he suggested pension and sovereign wealth funds buying stakes in AI companies. Queens' College Cambridge · c. Mar 2025 · 0:18:29 ↗transcript

Elon Musk (xAI), 2023–26, Demis Hassabis (DeepMind), Apr 2025, and Greg Brockman (OpenAI), Mar 2019: yes, it is the rare positive picture.

"The Banks culture books are probably the best envisioning, in fact not probably, they're definitely by far the best envisioning of an AI future. There's nothing even close."

Musk, unprompted, straight after predicting "universal high income". He has raised the books himself at least five times since, calling them a picture of a probable positive AI future and of what he thinks the future will probably be like. Rishi Sunak conversation · Nov 2023 · 0:17:18 ↗transcript

"And the best envisioning of an AI future that I've read are the books by Ian Banks, the culture books."

Musk, on how society gets through job losses. The Economist's editor-in-chief replies that she is reading the books because of him: "But I actually don't find that future a terribly appealing one because I don't think humans have agency in that future." The unlabelled replies that follow, "They have agency in their small way" and that Banks "had a very different political philosophy", are most likely Musk's. The Economist · Jul 2026 · 0:33:46 ↗transcript

"So if you want an optimistic thing, I would say the Culture series by Ian Banks. I think that's the best depiction of a post-AGI universe where, you know, AIs and you've basically got societies of AIs and humans and kind of alien species, actually, and sort of maximum human flourishing across the galaxy."

Hassabis, asked for a book that makes him optimistic. He has named the Culture as his model future again and again, as recently as Oct 2026. Possible · Apr 2025 · 0:51:11 ↗transcript

"I think that for AGI, there's actually very few positive stories out there. One that I think is there is Banks' culture series."

Brockman, asked what will be scarce once AGI succeeds. His point is that even in a post-scarcity world, finding meaning takes effort. Greymatter · Mar 2019 · 0:28:02 ↗transcript

Mustafa Suleyman (Microsoft AI), 2026: no. AI should serve the 8 billion people on Earth, not go to the stars, and he names Musk.

"That humans remain at the top of the food chain. … And, you know, some of the things that you hear from Elon often, or even others in the field. You can clearly see that they're sort of fixating on a world in 2050 or 2075 when they're going off and exploring other universes and conquering resources from other planets and stuff like that. And a system like that is unclear to me how it would have any time for preserving us as a species."

Suleyman, on building only AI that stays subordinate to us. Financial Times · Feb 2026 · 0:07:11 ↗transcript

"The idea that we're going to unleash this new species that is somehow going to go and explore the galaxies and leave humans behind, I think that most of the humans on the planet today would have some problems with that idea."

Arguing that AI should deliver healthcare and education first. In the same interview he recalls Larry Page telling him he would explore the galaxies at the head of a von Neumann probe, and he defers such probes and mind uploading by many centuries. The Builders · Oct 2026 · 0:09:31 ↗transcript