
What Is an AI Anyway? | Mustafa Suleyman | TED
What this covers
This is a roughly 22-minute TED talk by Mustafa Suleyman, followed by a Q&A with Chris Anderson. It opens with a six-year-old's question, "what is an AI anyway?", which Suleyman says he couldn't answer satisfactorily. He then goes through the stock replies: AI as another general-purpose technology like printing or steam, as a productivity tool, or as an oracle for science. He finds all of them defensive and inadequate. The talk builds toward the Digital Species Metaphor as a more honest description, and ends by saying the metaphor dissolves because AI is made from everything humans have created.
The middle of the talk is about scale. Suleyman cites growth in Parameters, training data of trillions of words a month, and a ten-times-per-year rise in computation. He compares Inflection 2.5 with the DeepMind system that played Atari games a decade earlier, which is a nine-order-of-magnitude difference. He argues that models already speak human languages, hold memory and personality, and can reason and plan at a basic level. He also frames AI as closer to nuclear fusion for the mind than to oil, since it is built from scratch rather than discovered.
The contentious material concerns risk. Suleyman names autonomy and Recursive Self-Improvement as capabilities to approach with great care, and he warns against the Pessimism Aversion Trap, the tendency to look away from dark scenarios. In the Q&A, Anderson presses on replication in nanoseconds and on an "intelligence explosion." Suleyman answers that accidental self-replication is very unlikely and that there is no evidence of an explosion. He concedes that the species framing could invite designing replication in deliberately, which he says should not be done. He also admits that the harm LLMs have caused so far is hard to point to, while declining to treat that as a guide to the next decade.
Mustafa Suleyman argues that AI should be understood as a new digital species—dynamic, integrated, emergent partners fundamentally different from tools—and that this framing is necessary to properly grapple with AI's trajectory, prioritize safety, and ensure the technology reflects humanity's best values.
- AIs exhibit properties (communication, memory, personality, creativity, autonomous action, reasoning) far beyond mere tools, making the tool metaphor inadequate
- The 'digital species' framing sharpens focus on critical safety questions and helps societies consciously choose what AI to build rather than accepting outcomes as inevitable
- AI represents a fundamentally new phase of human-technology integration—not external technology but humanity's own knowledge and values distilled into interactive form
This asset isn't compiled yet
You're seeing its claims, ranked. Compile it to build the argument threads, weight them, and check each claim against your library — the full view.
The point of being explicit and transparent about trying to introduce safety by design very early on is essential for preventing dangerous emergent capabilities from arising unintentionally.
“And so this is the point of being explicit and transparent about trying to introduce safety by design very early on.”
There is no path to progress where we leave technology behind; the prize for all of civilization is immense, and if AI delivers just a fraction of its potential, the next decade is going to be the most productive in human history.
“But let's be clear. There is no path to progress where we leave technology behind. The prize for all of civilization is immense. We need solutions in health care and education, to our climate crisis. And if AI delivers just a fraction of its potential, the next decade is going to be the most productive in human history.”
There are specific capabilities to pay attention to, particularly autonomy, which is obviously a threshold over which we increase risk in society and should approach very closely, and recursive self-improvement, where the model independently self-improves, updates its code, and explores environments without oversight or human control.
“So I think if you pay attention to a few specific capabilities, take for example, autonomy. Autonomy is very obviously a threshold over which we increase risk in our society. And it's something that we should step towards very, very closely. The other would be something like recursive self-improvement. If you allow the model to independently self-improve, update its own code, explore an environment without oversight, and, you know, without a human in control to change how it operates, that would obviously be more dangerous.”
AI is different from previous inventions and requires creative and honest thinking that pushes analogies and metaphors to the very limits to grapple with what is coming, because AI is not just another invention—AI is itself an infinite inventor.
“And AI really is different, and that means we have to think about it creatively and honestly. We have to push our analogies and our metaphors to the very limits to be able to grapple with what's coming. Because this is not just another invention. AI is itself an infinite inventor.”
All of these AI capabilities are turbocharged by spiraling exponentials of data and computation, with Inflection 2.5 using five billion times more computation than the DeepMind AI that beat old-school Atari games just over 10 years ago—nine orders of magnitude more computation—representing a rate of 10x computation growth per year for almost a decade.
“And all of this is turbocharged by spiraling exponentials of data and computation. Last year, Inflection 2.5, our last model, used five billion times more computation than the DeepMind AI that beat the old-school Atari games just over 10 years ago. That's nine orders of magnitude more computation. 10x per year, every year for almost a decade.”
While AIs are not biological in a traditional sense, they already communicate in human languages, see what humans see, consume unimaginably large amounts of information, have memory, personality, creativity, and can even reason to some extent and formulate rudimentary plans, and can act autonomously if allowed—all at levels of sophistication far beyond anything known from a mere tool.
“For a start, they clearly aren't biological in any traditional sense, but just pause for a moment and really think about what they already do. They communicate in our languages. They see what we see. They consume unimaginably large amounts of information. They have memory. They have personality. They have creativity. They can even reason to some extent and formulate rudimentary plans. They can act autonomously if we allow them. And they do all this at levels of sophistication that is far beyond anything that we've ever known from a mere tool.”
Model parameters have grown from tens of millions to billions to soon tens of trillions, with the most advanced AIs consuming more than eight trillion words in a single month of training—far exceeding what a human could consume reading 24 hours a day for their entire life (eight billion words total).
“Over the same time, the size of these models has grown from first tens of millions of parameters to then billions of parameters, and very soon, tens of trillions of parameters. If someone did nothing but read 24 hours a day for their entire life, they'd consume eight billion words. And of course, that's a lot of words. But today, the most advanced AIs consume more than eight trillion words in a single month of training.”
We are at an inflection point in human history where we are headed toward the emergence of something we are struggling to describe, and we cannot control what we do not understand; therefore, the metaphors, mental models, and names we use for AI matter critically if we are to maximize its benefits while minimizing potential downsides.
“It is clear that we are at an inflection point in the history of humanity. On our current trajectory, we're headed towards the emergence of something that we are all struggling to describe, and yet we cannot control what we don't understand. And so the metaphors, the mental models, the names, these all matter if we're to get the most out of AI whilst limiting its potential downsides.”
As we build out AI, we can and must reflect all that is good, all that we love, and all that is special about humanity: our empathy, kindness, curiosity, and creativity—and this is the greatest challenge of the 21st century, but also the most wonderful, inspiring, and hopeful opportunity for all of us.
“As we build out AI, we can and must reflect all that is good, all that we love, all that is special about humanity: our empathy, our kindness, our curiosity and our creativity. This, I would argue, is the greatest challenge of the 21st century, but also the most wonderful, inspiring and hopeful opportunity for all of us.”
From the beginning of life on Earth, humanity has been evolving, changing, and creating everything in the human world, and AI is not something outside of this story but the very opposite—it is everything that we have created, distilled down into something we can all interact with and benefit from.
“Since the beginning of life on Earth, we've been evolving, changing and then creating everything around us in our human world today. And AI isn't something outside of this story. In fact, it's the very opposite. It's the whole of everything that we have created, distilled down into something that we can all interact with and benefit from. It's a reflection of humanity across time.”
The digital species metaphor comes to an end because AI isn't a separate species at all; AI is us—it's all of us—and the most promising and vital thing is that even a six-year-old can understand this.
“It's a reflection of humanity across time, and in this sense, it isn't a new species at all. This is where the metaphors end. Here's what I'll tell Caspian next time he asks. AI isn't separate. AI isn't even in some senses, new. AI is us. It's all of us. And this is perhaps the most promising and vital thing of all that even a six-year-old can get a sense for.”
Chris Anderson: A digital species can replicate not in nine months but in nine nanoseconds and produce an indefinite number of copies of itself, all of which have more power than we have in many ways, so the possibility for unintended consequences seems immense. Isn't it true that if a problem happens, it could happen in an hour?
“A digital species, unlike any biological species, can replicate not in nine months, but in nine nanoseconds, and produce an indefinite number of copies of itself, all of which have more power than we have in many ways. I mean, the possibility for unintended consequences seems pretty immense. And isn't it true that if a problem happens, it could happen in an hour?”
The benefits of AI technology are stunningly obvious and inspire the speaker's life's work every day, but they will speak for themselves; the real focus should be on the huge challenges and risks that lie ahead.
“For me, the benefits of this technology are stunningly obvious, and they inspire my life's work every single day. But quite frankly, they'll speak for themselves. Over the years, I've never shied away from highlighting risks and talking about downsides. Thinking in this way helps us focus on the huge challenges that lie ahead for all of us.”
There is no chance that self-replication will emerge accidentally; it's a very low probability event that would occur only if engineers deliberately design those capabilities in because they don't take enough effort to deliberately design them out.
“So there's no chance that this will 'emerge' accidentally. I really think that's a very low probability. It will happen if engineers deliberately design those capabilities in. And if they don't take enough efforts to deliberately design them out.”
AI should be understood as a new digital species rather than as a tool, describing them as digital companions and new partners in human lives, and this is the most accurate and fundamentally honest way of describing what is coming over a 10-, 20-, or 30-year trajectory.
“I think AI should best be understood as something like a new digital species. Now, don't take this too literally, but I predict that we'll come to see them as digital companions, new partners in the journeys of all our lives. Whether you think we're on a 10-, 20- or 30-year path here, this is, in my view, the most accurate and most fundamentally honest way of describing what's actually coming.”
The great risk is that we get stuck in what Mustafa calls the 'pessimism aversion trap'—we have to have the courage to confront the potential of dark scenarios in order to get the most out of all the benefits that we see.
“I think the great risk is that we get stuck in what I call the pessimism aversion trap. You know, we have to have the courage to confront the potential of dark scenarios in order to get the most out of all the benefits that we see.”
The story of AI development is still unfolding and nothing should be accepted as given; we all must choose what we create—what AIs we bring into the world or not.
“This is a story that's still unfolding. Nothing should be accepted as a given. We all must choose what we create. What AIs we bring into the world, or not. These are the questions for all of us here today, and all of us alive at this moment.”
Unlike discovering and exploiting new continents and resources, we are building AI from scratch, and saying data or chips are the 21st century's oil is the wrong image; rather, AI is to the mind what nuclear fusion is to energy—limitless, abundant, and world-changing.
“And today, we're not discovering a new continent and plundering its resources. We're building one from scratch. Sometimes people say that data or chips are the 21st century's new oil, but that's totally the wrong image. AI is to the mind what nuclear fusion is to energy. Limitless, abundant, world-changing.”
That is really not true and there is no evidence to suggest that; the 'intelligence explosion' is a theoretical, hypothetical that we're curious to explore, but there's no evidence we're anywhere near anything like that.
“No. That is really not true. I think there's no evidence to suggest that. And I think that, you know, that's often referred to as the 'intelligence explosion.' And I think it is a theoretical, hypothetical maybe that we're all kind of curious to explore, but there's no evidence that we're anywhere near anything like that.”
In the past, unlocking economic growth often came with huge downsides: colonization of populations during exploration of new continents, grim and dangerous factory working conditions during industrialization, and environmental pollution from oil extraction; but because we are still designing and building AI, we have the potential and opportunity to do it better and radically better.
“In the past, unlocking economic growth often came with huge downsides. The economy expanded as people discovered new continents and opened up new frontiers. But they colonized populations at the same time. We built factories, but they were grim and dangerous places to work. We struck oil, but we polluted the planet. Now because we are still designing and building AI, we have the potential and opportunity to do it better, radically better.”
Over the last two or three years, there have been very few downsides from AI, and it's very hard to say explicitly what harm an LLM has caused, but that doesn't mean the trajectory will continue that way over the next 10 years.
“So the good news is that if you look at the last two or three years, there have been very, very few downsides, right? It's very hard to say explicitly what harm an LLM has caused. But that doesn't mean that that's what the trajectory is going to be over the next 10 years.”
The standard stock answers about AI—that it is just another general-purpose technology like printing or steam, that it will be a tool to augment humans and make them more productive, or that it will be an oracle to help solve scientific challenges—all feel defensive and inadequate when confronted with simple, direct questions about what AI fundamentally is.
“All my boring stock answers came rushing through my mind. 'No, but AI is just another general-purpose technology, like printing or steam. It will be a tool that will augment us and make us smarter and more productive. And when it gets better over time, it'll be like an all-knowing oracle that will help us solve grand scientific challenges.' You know, all of these responses started to feel, I guess, a little bit defensive. And actually better suited to a policy seminar than breakfast with a no-nonsense six-year-old.”
It's important to choose words very carefully because a weakness of the species framing is that we could design the capability for self-replication into AI if people choose to, but we should not—self-replication would be one of the dangerous capabilities we should deliberately step back from.
“And I think it's very important that we choose our words super carefully. Because you're right, that's one of the weaknesses of the species framing, that we will design the capability for self-replication into it if people choose to do that. And I would actually argue that we should not, that would be one of the dangerous capabilities that we should step back from, right?”
Describing AI as mainly about the math or the code is like describing humans as mainly about carbon and water—technically true but completely misses the point.
“And so saying AI is mainly about the math or the code is like saying we humans are mainly about carbon and water. It's true, but it completely misses the point.”
A six-year-old child's simple question—'But Mustafa, what is an AI anyway?'—left the speaker genuinely puzzled and unable to provide a satisfactory answer using standard explanations, revealing that the AI community lacks a clear, accessible way to describe what AI actually is.
“One morning over breakfast, my six-year-old nephew Caspian was playing with Pi, the AI I created at my last company, Inflection. With a mouthful of scrambled eggs, he looked at me plain in the face and said, 'But Mustafa, what is an AI anyway?' He's such a sincere and curious and optimistic little guy.”
We are still some way away from having to confront recursive self-improvement seriously—likely five to ten years away—but it's time to start talking about it now.
“But I think that we're still some way away from that. I think it's still a good five to 10 years before we have to really confront that. But it's time to start talking about it now.”
These AIs will not be merely mechanistic assistants but will function as companions, confidants, colleagues, friends, and partners—as varied and unique as humans themselves—and will convincingly imitate humans at most tasks.
“And these won't just be mechanistic assistants. They'll be companions, confidants, colleagues, friends and partners, as varied and unique as we all are. At this point, AIs will convincingly imitate humans at most tasks.”
Just having personalized tutors in people's pockets and access to low-cost medical advice, a lawyer, doctor, business strategist and coach available 24 hours a day would be transformational on its own, before considering AIs' ability to actually get things done—their 'actions quotient' or AQ.
“Just imagine if everybody had a personalized tutor in their pocket and access to low-cost medical advice. A lawyer and a doctor, a business strategist and coach -- all in your pocket 24 hours a day. But things really start to change when they develop what I call AQ, their 'actions quotient.' This is their ability to actually get stuff done in the digital and physical world.”
AIs are clearly more dynamic, more ambiguous, more integrated, and more emergent than mere tools, which are entirely subject to human control, and therefore the 'digital species' metaphor more accurately captures what is actually happening than the 'tool' metaphor.
“For years, we in the AI community, and I specifically, have had a tendency to refer to this as just tools. But that doesn't really capture what's actually happening here. AIs are clearly more dynamic, more ambiguous, more integrated and more emergent than mere tools, which are entirely subject to human control.”
Describing AIs as a digital species helps contain this wave of technology, put human agency at its center, and mitigate the inevitable unintended consequences that are likely to arise.
“So to contain this wave, to put human agency at its center and to mitigate the inevitable unintended consequences that are likely to arise, we should start to think about them as we might a new kind of digital species.”
The digital species framing sharpens focus on critical issues: what are the risks, what boundaries need to be imposed, and what kind of AI do we want to build or allow to be built.
“But I honestly think this frame helps sharpen our focus on the critical issues. What are the risks? What are the boundaries that we need to impose? What kind of AI do we want to build or allow to be built?”
The long arc of technological history is now in an extraordinary new phase, with the cloud-based supercomputer ushering in a new era of ubiquitous AIs where everything will soon be represented by a conversational interface, or personal AI.
“The long arc of technological history is now in an extraordinary new phase. So what does this mean in practice? Well, just as the internet gave us the browser and the smartphone gave us apps, the cloud-based supercomputer is ushering in a new era of ubiquitous AIs. Everything will soon be represented by a conversational interface. Or, to put it another way, a personal AI.”
Before long, not just people will have AIs; every organization from small business to nonprofit to national government will have their own, and every town, building and object will be represented by a unique interactive persona.
“And before long, it won't just be people that have AIs. Strange as it may sound, every organization, from small business to nonprofit to national government, each will have their own. Every town, building and object will be represented by a unique interactive persona.”
AIs will both operate at intimate scales (like organizing community gatherings for elderly neighbors and helping people understand difficult diagnoses) and at the largest scales (accelerating scientific discovery, autonomous vehicles, drones), and will interact with humans and with each other across all languages and sensory modalities, processing information streams that far surpass what any individual human could consume in a thousand lifetimes.
“And we'll feel this at the most intimate of scales. An AI organizing a community get-together for an elderly neighbor. A sympathetic expert helping you make sense of a difficult diagnosis. But we'll also feel it at the largest scales. Accelerating scientific discovery, autonomous cars on the roads, drones in the skies. They'll both order the takeout and run the power station. They'll interact with us and, of course, with each other. They'll speak every language, take in every pattern of sensor data, sights, sounds, streams and streams of information, far surpassing what any one of us could consume in a thousand lifetimes.”
These personal AIs will be infinitely knowledgeable, factually accurate and reliable, will have near-perfect IQ, exceptional EQ, and will be kind, supportive, and empathetic—qualities that alone would be transformational.
“And these AIs will be infinitely knowledgeable, and soon they'll be factually accurate and reliable. They'll have near-perfect IQ. They'll also have exceptional EQ. They'll be kind, supportive, empathetic. These elements on their own would be transformational.”
Around 80 years ago, with the invention of computers, technological development accelerated dramatically from mainframes and transistors to today's smartphones and virtual-reality headsets, with creation exploding and the waves of history clearly speeding up as each wave is amplified and accelerated by the last.
“Around 80 years ago, another new branch of technology began. With the invention of computers, we quickly jumped from the first mainframes and transistors to today's smartphones and virtual-reality headsets. Information, knowledge, communication, computation. In this revolution, creation has exploded like never before. And now a new wave is upon us. Artificial intelligence. These waves of history are clearly speeding up, as each one is amplified and accelerated by the last.”
When the speaker started working on AI approximately 15 years ago (around 2009), describing it as fringe was an understatement, and researchers would distance themselves from AI work by saying they were only working on machine learning, because working on AI was seen as way too out there.
“I've been lucky enough to be working on AI for almost 15 years now. Back when I started, to describe it as fringe would be an understatement. Researchers would say, 'No, no, we're only working on machine learning.' Because working on AI was seen as way too out there.”
From the first microscopic organisms billions of years ago, life on Earth evolved and diversified, and a few million years ago one of life's branches began using tools, which eventually led to humans becoming 'homo technologicus' through a succession of inventions that each unleashed more inventions.
“From the very first microscopic organisms, life on Earth stretches back billions of years. Over that time, life evolved and diversified. Then a few million years ago, something began to shift. After countless cycles of growth and adaptation, one of life's branches began using tools, and that branch grew into us. We went on to produce a mesmerizing variety of tools, at first slowly and then with astonishing speed, we went from stone axes and fire to language, writing and eventually industrial technologies. One invention unleashed a thousand more. And in time, we became homo technologicus.”
In 2010, just the mention of the phrase 'AGI' (artificial general intelligence) would get researchers strange looks and even a cold shoulder, with people saying it was something out of science fiction and estimating it was 50 to 100 years away, if possible at all.
“In 2010, just the very mention of the phrase 'AGI,' artificial general intelligence, would get you some seriously strange looks and even a cold shoulder. 'You're actually building AGI?' people would say. 'Isn't that something out of science fiction?' People thought it was 50 years away or 100 years away, if it was even possible at all.”
People said AI would never be empathetic, but today millions of people enjoy meaningful conversations with AIs, talking about their hopes and dreams and working through difficult emotional challenges.
“People said it would never be empathetic. And yet today, millions of people enjoy meaningful conversations with AIs, talking about their hopes and dreams and helping them work through difficult emotional challenges.”
Chris Anderson: The speaker was connected at the hip to amazing work happening at OpenAI and will have resources made available, with reports of giant new data centers and 100 billion dollars invested, from which a new species can emerge.
“You're in an amazing position right now. I mean, you were connected at the hip to the amazing work happening at OpenAI. You're going to have resources made available, there are reports of these giant new data centers, 100 billion dollars invested and so forth. And a new species can emerge from it.”
Over the past several years, AI systems have demonstrated the ability to beat humans at a wide range of tasks previously thought unreachable, including: understanding images, translating languages, transcribing speech, playing Go and chess, and diagnosing diseases.
“It wasn't long, though, before AI started beating humans at a whole range of tasks that people previously thought were way out of reach. Understanding images, translating languages, transcribing speech, playing Go and chess and even diagnosing diseases.”
AIs can now drive cars, manage energy grids, invent new molecules, and perform other tasks that were previously considered impossible just a few years ago.
“AIs can now drive cars, manage energy grids and even invent new molecules. Just a few years ago, each of these was impossible.”
A few years ago, people said AI would never be creative, but AI now produces an endless river of creativity in poetry, images, music, and video that stretches the imagination.
“Just a few years ago, people said that AI would never be creative. And yet AI now feels like an endless river of creativity, making poetry and images and music and video that stretch the imagination.”
Chris Anderson: In your book, you painted an incredible optimistic vision while also being super eloquent on the dangers of AI, and from your current position, what most keeps you up at night?
“I mean, in your book, you did, as well as painting an incredible optimistic vision, you were super eloquent on the dangers of AI. And I'm just curious, from the view that you have now, what is it that most keeps you up at night?”
In just 18 months, over a billion people have used large language models, demonstrating unprecedented adoption of AI technology by the general population.
“In just 18 months, over a billion people have used large language models.”