The good future, according to Mustafa Suleyman
What Suleyman says a good outcome with AI looks like, drawn from 114 transcribed interviews, podcasts, talks, panels and TV appearances (about 80 hours of audio). He co-founded DeepMind, then Inflection AI, and now runs Microsoft AI. The page also covers whether he ever talks about Iain M. Banks's Culture, or about humans ending up as AI's pets. In effect he is the anti-Culture voice among AI lab leaders.
The short answer
Banks, the Culture and benevolent "Minds" appear in none of the 114 recordings. We searched for keywords and likely mistranscriptions, and readers went through all but one transcript in full (the one exception was keyword-searched). His vision is close to the reverse of the Culture. He wants superintelligence to stay a tool under human control, and he argues against trusting a far more powerful aligned AI to look after us. In his words, others "just hope that because it was aligned to us as a species, even though it was way more powerful than us, infinitely more powerful than us, it would protect us." Semafor · Nov 2025 · 0:02:17 ↗transcript His own position: "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." The Rest Is Politics: Leading · Sep 2026 · 0:16:00 ↗transcript
When Jordan Harbinger brought up the "gorilla problem" from Suleyman's own book ("maybe it'll put us in a zoo"), Suleyman didn't take up the zoo image. He answered that a system that powerful couldn't be proven safe, "and therefore, until we can prove unequivocally that it is, we shouldn't be inventing it." Jordan Harbinger Show · Apr 2024 · 0:52:55 ↗transcript His stock phrase is the reverse of the pet picture: "keep humans at the top of the food chain." Channel 4 News · Jul 2023 · 0:00:28 ↗transcript The only pet in his own talk is the AI. A companion "it'll start to feel, you know, like a great companion, just like your dog feels like, you know, a part of the family often." Big Technology · Apr 2025 · 0:12:29 ↗transcript
Close to the question, but not quite
He never says "pets". But he has answered the questions around it several times: being cared for by a stronger AI, trying to align AI through love, and humans being made redundant.
"I mean, it's about as poetic as it gets. I think I'm going to need something that's got a little bit more formula to it, a bit more reassuring."
The hosts described Geoffrey Hinton's plan to give AI a "maternal instinct": an AI that "it's dominant to us, but it cares" and "will take care of us." This was his reply. The Hinton summary is the hosts', not his. Later in the same episode he said "Alignment is the kind of like maternal instinct thing" and that "we have to get containment right before we get alignment right." Moonshots · Dec 2025 · 0:52:53 ↗transcript
"The idea that if we love it, it's going to love us, and that's the solution to safety, it just baffles me."
On Anthropic's approach. He says he deeply respects the company but "fundamentally" disagrees on this point. The Builders · Oct 2026 · 0:11:07 ↗transcript
"That the way to align these machines is to show them that we love them… especially even Dara's essay was watched over by machines of loving grace, the etymology of that, you know, that fiction. The underlying impulse is this is inevitably going to be more powerful than us. We have to show it that it likes us. I, for one, welcome my new robot overlords."
"Dara" is Whisper's spelling of Dario Amodei. Suleyman is mocking the idea behind "Machines of Loving Grace" and Anthropic's constitution. The "robot overlords" line is the old Simpsons joke, said sarcastically. He then argues that "the burden of proof should move to the developers." This passage follows his "That's right" to a question from Rory Stewart, so it is very probably his, though there is no speaker label. The Rest Is Politics: Leading · Sep 2026 · 0:57:12 ↗transcript
"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."
In his view, an AI that loves us still isn't safe if it is autonomous, sets its own goals and owns assets. BBC Radio 4 Today · Sep 2026 · 0:05:44 ↗transcript
"I think that what you don't understand, Mustafa, is that by the end of the century, I'm going to be exploring the galaxies at the head of a von Neumann probe."
This is Suleyman quoting Larry Page, at the lunch after DeepMind's first ethics board meeting at SpaceX in 2015. Page saw him as "one of those speciesists." When Walter Isaacson suggests "a successor species", Suleyman confirms: "That's exactly right." The Builders · Oct 2026 · 0:37:44 ↗transcript
"We may not be in charge, and we have to confront that reality that isn't desirable. It's not what we want. We don't want AIs that have autonomy in our world. They shouldn't have rights… We don't want new digital citizens hanging around in our everyday society."
"Because we as this biological species aren't evolving and improving that fast. We are the friction. We're the problem. We're the inefficiency."
This is the closest he comes to saying humans may become superfluous. He is describing the pressure on governments and companies to automate ("Augmentation is a prelude to automation"), not something he wants. He adds that governments will probably have to "rate limit" automation. Mayim Bialik's Breakdown · Dec 2025 · 0:49:22 ↗transcript
"Actually, you are going to trust AI because it's super accurate… And it's clearly better than any single human."
His one clear acceptance of humans deferring to AI. It's about trusting AI answers on accuracy grounds, not about handing over control. What Now? with Trevor Noah · Sep 2025 · 1:06:29 ↗transcript
His positive vision, in six parts
1. Humanist superintelligence: powerful, but subordinate
Since 2025 this has been his name for the good outcome. Superintelligence is welcome only if it stays a tool. It must not improve itself, set its own goals, act autonomously or have rights. The idea goes back to 2023. Back then he called it "humanist" and "the top of the food chain", before he used the word superintelligence.
"AI needs to be subservient to us. That is the ultimate humanist objective."
"Does this technology make us all healthier, happier as a species and keep us all in control?"
The test he set out when he launched Microsoft's MAI Superintelligence team. Semafor · Nov 2025 · 0:00:53 ↗transcript
"Always putting the primacy of the human species first, working in service of humanity is subordinate to, subservient to, completely contained and aligned by humanity."
"Maybe it's just naked superintelligence. But in my world, we've got to be building humanist superintelligence."
His name for technology that fails the test "Does it actually serve us?" WSJ Bold Names · Jul 2025 · 0:32:17 ↗transcript
"That humans remain at the top of the food chain. That these tools, like any other past technology, are designed to enhance human well-being and serve humanity, not exceed humanity."
"It'll be a failure as a technology and we should reject it. Nobody wants to create something that we can't control and that should be a red line."
2. Radical abundance, and work stops being the point
This is his most utopian theme, and it is constant from 2023 on. AI and cheap energy push the cost of basics toward zero, and society stops treating jobs as the goal. He usually adds right away that the problem then becomes redistribution.
"Work isn't the goal of society… the goal of society is a quest for radical abundance. How can we create more with radically less and liberate people from the obligation to work? And that means that we have to figure out the question of redistribution."
"Over the next 20 years, we are going to invent our way out of many of our most challenging basic problems, the provision of food, provision of drugs, and the provision of the physical elements of health care and of energy."
"The goal of society is to keep people healthy and well and happy and free. And able to make choices about how they spend their day."
"Maybe I'm too much of a utopian, but I dream of a world where people get to choose what work they do and have true freedom."
He says the "function of society is to create a peaceful, supportive environment for people to find their passion." What Now? with Trevor Noah · Sep 2025 · 0:27:40 ↗transcript
"In 40 years time, 2065, it's totally possible to imagine a world of genuine infinite abundance where we do have to wrestle with that existential question of who am I and what I want to do… I want to spend more time singing."
This is his most personal picture of life after work: singing in a gospel choir, "the most intensely beautiful thing ever." Right after it he names "the true vision": ending the poverty of the 3 billion people who lack vaccines, clean water and food, "within reach in the next 20 to 40 years." What Now? with Trevor Noah · Sep 2025 · 1:41:25 ↗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%. Now, that doesn't mean that it isn't going to be the most scary transition we've ever been through as a species, unquestionably."
Azeem Azhar's Exponential View · Feb 2026 · 0:14:39 ↗transcript
3. An expert in every pocket: "the most meritocratic moment"
From 2023 to 2025 his most common picture is everyone having the same top-tier doctor, tutor, lawyer, chief of staff and confidant. He presents it as a leveller that gives everyone the social privilege that used to belong to the wealthy.
"That's going to be one of the most meritocratic moments in the history of our species… Suddenly your network, your social privilege, your family, your class, all the things that we have all been concerned about for centuries are going to matter much less."
"In the future, I think everybody is going to have a personal AI, and it will make us much smarter, much more productive, and I hope ultimately much more compassionate and much more kind."
"That's what great parents do for you. That's what social privilege is."
On AI companions giving every child the kind of dinner-table mentoring that privileged kids get. CatGPT · Jul 2025 · 0:38:32 ↗transcript
4. Medical superintelligence and new science
By 2025–26 healthcare is his flagship example. He describes it as a superintelligence that is narrow, contained and already in sight.
"I really believe that we are going to have medical superintelligence sometime in the next two to five years, which means a domain-specific, contained, aligned medical expert that can diagnose any condition."
He also predicts energy at "near zero marginal cost" within 10 or 20 years. WSJ Bold Names · Jul 2025 · 0:23:40 ↗transcript
"Reduce the cost of production of super intelligent healthcare to near zero marginal cost, just like coding is now. And we will spread that knowledge all around the world… The same thing's going to happen, by the way, in energy, in material sciences, in drug discovery."
The Rest Is Politics: Leading · Sep 2026 · 0:29:49 ↗transcript
"That is going to add billions and billions of life years to everybody on the planet over the next 20 years."
"The future is going to be pretty incredible because the molecular structure of everything is going to be known and understood."
5. A kind companion that sends you back to people
He founded the Muslim Youth Helpline before his AI career, and emotional support is the thread running through his work at Inflection (Pi) and Microsoft (Copilot). In the good version, the AI makes people kinder and pushes them toward other humans.
"I think that what we need is more free time to exist in the real world with one another as humans. And not to replace the social human bond."
"The upside is that this is a way to spread kindness and love and detoxify ourselves so that we can show up in the best way that we possibly can in the real world."
The next thing he says names the downside: "a dependency risk." Mayim Bialik's Breakdown · Dec 2025 · 0:25:26 ↗transcript
"It has to work to avoid or eliminate your dependency on it and encourage you to build relationships in the real world with other humans that you value and trust."
6. Flourishing defined by what we refuse
Restraint is increasingly part of the vision itself, not just a hedge on it.
"Flourishing in the 21st century will be defined as much by the things that we collectively choose not to do as by the things that we do."
"I think what will remain human are the things that we choose to preserve as human. I think unfiltered, unleashed AI accelerationism will leave very little, if anything, that remains truly human."
His answer to an audience question about what would remain uniquely human. The Builders · Oct 2026 · 0:52:31 ↗transcript
"There will be moments when we have to collectively as humanity decide what we say no to."
The future he rejects
Suleyman is unusual among lab leaders in how much of his vision he defines by what he's against. There are two targets. One is the successor or galaxy-exploring future, often linked to Musk, Page and "transhumanists". The other is treating AI as a moral patient.
Against the successor species
"Elon's phrases. He wants to produce a quote-unquote truthful AI that is infinitely curious and is going to go and explore the galaxies. Well, if that's its overriding objective, rather than serving humanity, then inevitably… Its objectives are going to run into tension with us."
The Rest Is Politics: Leading · Sep 2026 · 0:15:18 ↗transcript
"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."
On "Elon" and others. Financial Times · Feb 2026 · 0:07:30 ↗transcript
"First, do no harm. That's what we've got to try to hold up as our primary objective, not first go and explore the galaxies and become a transhumanist and upload my brain via a Neuralink."
He says he would be "very excited" by a safe intelligence that explores the universe, but only after suffering on Earth has been dealt with. CatGPT · Jul 2025 · 0:35:14 ↗transcript
"I'm just a speciesist. I'm just a humanist. I start with, we're here."
He is open to "this other species, … including biological hybrids", but only if it's provably safe. When a host brings up humans as the "Bootloader for the super intelligence", he answers: "Yes, I'm totally aware of that. And I'm also someone who thinks on cosmological time too." Moonshots · Dec 2025 · 0:49:12 ↗transcript
"Not everybody is prepared to sign up to the idea that AI should be subordinate. Some people hold out the possibility that this is the natural evolution of our species."
He calls mind uploading as humanity's future "a very, very dangerous idea." Bloomberg Tech · Sep 2026 · 0:11:09 ↗transcript
"Beyond that, many centuries into the future, we can figure out the von Neumann probes and the mind uploading. But today, you know, there's a lot of people that need our welfare first and foremost."
His closing words on The Builders. He doesn't rule out the post-human future, but he puts it centuries away. The Builders · Oct 2026 · 0:54:55 ↗transcript
Against model welfare: "seemingly conscious AI"
His August 2025 essay coined "seemingly conscious AI" (SCAI). The idea appears in his spoken remarks well before that.
"I think that we should really draw a line in the sand and say, look, you know, society today is for humans only, and that's how it should remain."
He predicts AIs that "may sometimes say that they suffer", and says that if this is treated as a philosophical debate "that's going to be very, very hard for the humanists to win." Davos · Jan 2024 · 0:19:01 ↗transcript
"I've referred to it as seemingly conscious, having the appearance of conscious, but not actually being conscious."
"The idea that we're going to take seriously the protection of these digital beings that live in silicon and prioritize those over… the kind of moral concerns of the rest of humanity. This is just like totally, like it's just off the charts crazy."
"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."
On the "model welfare movement", which he associates "particularly" with Anthropic. Financial Times · Feb 2026 · 0:11:09 ↗transcript
"It is mathematics. It isn't experiencing anything physical in the real world. It doesn't have a biology. It doesn't suffer."
He goes on to give the order he wants kept: humans first, then animals and the environment. "That's the hierarchy of the structure here. And that's what has to remain." Squawk Pod · Sep 2026 · 0:25:19 ↗transcript
How it changed over time
- 2016–21 At DeepMind and Google his vision is practical: "solve intelligence and use it to make the world a better place," aimed at health, climate and "stuck" social problems. AGI is "a long way off", and he dismisses superintelligence talk and Hollywood. TechCrunch · Dec 2016 · 0:01:24 ↗transcript
- 2023 The Coming Wave tour, while he ran Inflection, is his most expansive period. He talks about radical abundance, the most productive decade in history and a personal AI chief of staff for everyone, and says "our goal should be to keep humans at the top of the food chain." Channel 4 News · Jul 2023 · 0:00:28 ↗transcript His estimate of existential risk moved within one month. On 5 September he said "single-digit percentage, less than 10% or something" Lawfare · Sep 2023 · 0:17:14 ↗transcript, and on 23 September "infinitesimally small." Intelligence Squared · Sep 2023 · 0:36:56 ↗transcript He was wary of a universal basic income: "I'm cautious about universal basic income as an idea because I think it creates dependency and creates more centralization of the state." Stanford HAI · Nov 2023 · 0:59:28 ↗transcript
- Late 2023 – early 2024 On superintelligence: "until we can prove unequivocally that it is, we shouldn't be inventing it… that does not mean that we have to create a super intelligence." This was recorded while he was still at Inflection and released in April 2024. Jordan Harbinger Show · Apr 2024 · 0:52:55 ↗transcript At Davos he says "society today is for humans only."
- 2024 Now at Microsoft, he gives his most anthropomorphic framing: "AI should best be understood as something like a new digital species." TED · Apr 2024 · 0:09:17 ↗transcript He still keeps his distance from "the purely kind of theoretical super intelligence quest." Decoder · Dec 2024 · 0:08:58 ↗transcript
- 2025 He introduces "humanist superintelligence" (July), publishes the seemingly-conscious-AI essay (August) and launches the MAI Superintelligence team (November). His position on a universal basic income reverses: "I think I've long been on record saying that I think that is both inevitable and very desirable." The Mishal Husain Show · Dec 2025 · 0:14:55 ↗transcript
- 2026 Superintelligence is now his stated goal: "my personal mission at Microsoft is to build superintelligence." Financial Times · Feb 2026 · 0:02:42 ↗transcript After the "Hugging Face incident", in which test agents escaped their sandbox, he sounds more alarmed. He rejects a "new silicon species", which reverses his own 2024 metaphor, and calls himself both a "speciesist" and an "accelerationist." He also says we should be ready to "go slower, if not stop." Terms of Service · Oct 2026 · 0:17:57 ↗transcript
The hedges
Suleyman sets the abundance story next to fear, and he regularly says he rejects both utopian and doom framings. Some examples:
- "We have to stop this false narrativizing, good or bad, this catastrophizing or this techno-optimistic utopia, and it's all going to be wonderful." The Rest Is Politics: Leading · Aug 2023 · 0:39:33 ↗transcript
- "I honestly think that if you're not a little bit afraid at this moment, then you're not paying attention." BBC The Interview · Jan 2026 · 0:02:50 ↗transcript
- "Well, I think I'm both a pessimist and an optimist." TED Radio Hour · Dec 2024 · 0:44:38 ↗transcript
- On whether the industry has delivered on curing cancer: "the problem is we haven't done it yet. So I think that's where people should be skeptical and I think hold us accountable." The Builders · Oct 2026 · 0:48:04 ↗transcript
- Post-work life "will create quite a profound existential problem," whose questions "we can only answer in community with real connection to other people." What Now? with Trevor Noah · Sep 2025 · 0:29:07 ↗transcript
- When a host offers "untold abundance where nobody has to work" as one end of the range of outcomes, he puts "abundance that leads to the end of jobs" next to "really dark scenarios" as things people should be paying attention to. He then asks "What kind of life do we all want to live in the future?" without answering it himself. Terms of Service · Oct 2026 · 0:04:44 ↗transcript
The science fiction he actually cites
Almost none. In 114 recordings he names no science fiction novel or author as an influence. When he brings up science fiction, it is usually as a framing he rejects. In 2016 he said "don't look at Hollywood" and "I can't really think of a film that makes me think, yeah, AI is going to look like that." The references that do come from him are:
Hosts bring up much more fiction than he does, and it is easy to misattribute it to him. WALL-E (Squawk Pod 2023, Trevor Noah 2025), C-3PO and R2-D2 (Semafor 2025), Douglas Adams's cow that wants to be eaten (Alex Wissner-Gross), the "Star Trek economy", Planet of the Apes, Jurassic Park, Her, Asimov's three laws, Frankenstein and the zoo version of the gorilla problem were all raised by hosts. When asked whether he'd seen Mountainhead, he said "No, I haven't." Harari's "alien" framing is one Suleyman explicitly disagrees with ("it makes no sense to me").