The TED AI Show: Could AI really achieve consciousness? w/ neuroscientist Anil Seth

[00:00:00] TED Audio Collective.

[00:00:35] You know this question. How long will it take before some massive breakthrough, some kind of singularity emerges and suddenly AI becomes self-aware? Before AI becomes conscious? But we're getting way ahead of ourselves.

[00:00:51] Lately, reports from AI researchers suggest that AI models are not improving at the same rate as before and are hitting the limits of so-called scaling laws, at least as far as pre-training is concerned.

[00:01:03] There's also worries that we're running out of useful data, that these systems require better quality and greater amounts of data to continue growing at this exponential pace.

[00:01:14] The road to a machine that can think for itself is long, and it's starting to sound like it may be even longer than we think.

[00:01:21] For now, clever interfaces like ChatGPT's advanced voice mode, the one I experimented with in an earlier episode this season, helps give some illusion of a human at the other end of this conversation with an AI.

[00:01:35] I was surprised by how much it actually delighted me, and even kind of tricked me, at least a tiny little bit, into feeling like ChatGPT was really listening. Like a friend would.

[00:01:46] The thing is, though, it's a slippery slope. We're building technology that is so good at emulating humans that we start ascribing human attributes to it.

[00:01:57] We start wondering, does this thing actually care? Is it actually conscious? And if not now, will it be at some point in the future?

[00:02:06] And by the way, what even is consciousness anyway?

[00:02:09] The answer is trickier than you might think. To unpack it, I spoke with someone who's been tackling this question from the inside out, from the perspective of the one thing we know is conscious, the human brain.

[00:02:26] One of my mentors, the philosopher Daniel Dennett, we sadly lost earlier this year.

[00:02:32] He said we should treat AI as tools rather than colleagues and always remember the difference.

[00:02:39] That's Anil Seth. He's a professor of cognitive and computational neuroscience at the University of Oxford.

[00:02:45] He studies human consciousness and wrote a great book about it. It's called Being You, A New Science of Consciousness.

[00:02:52] And that quote from his mentor, it's something that sticks with him.

[00:02:55] It sticks with me because I think we have this tendency to always project too much of ourselves into the technologies we build.

[00:03:04] I think this has been something humans have done over history.

[00:03:07] And it's always got into trouble because we tend to misunderstand then the capabilities of the machines we build.

[00:03:16] And also, we tend to diminish ourselves in the process.

[00:03:20] And I think the recent explosion of interest in AI is a very prominent example of how we've fallen prior to this problem at this moment.

[00:03:29] So this is why Anil's on the show today.

[00:03:32] He's come to share why he thinks it's imperative we see AI as a tool, not as a friend.

[00:03:37] And why that difference matters to not only the future of this technology, but also the future of human consciousness.

[00:03:47] I'm Bilal Siddhu, and this is the TEDAI Show, where we figure out how to live and thrive in a world where AI is changing everything.

[00:04:02] How will humans and machines work together in the future?

[00:04:05] We spend so much time discussing how the world's changing.

[00:04:10] It would be absolutely absurd to believe the role of the CEO is not.

[00:04:14] This is Imagine This, a podcast from BCG that helps CEOs consider possible futures for our world and their businesses.

[00:04:23] Listen wherever you get your podcasts.

[00:04:26] Add a little curiosity into your routine with TED Talks Daily, the podcast that brings you a new TED Talk every weekday.

[00:04:36] In less than 15 minutes a day, you'll go beyond the headlines and learn about the big ideas shaping your future.

[00:04:43] Coming up, how AI will change the way we communicate, how to be a better leader, and more.

[00:04:49] Listen to TED Talks Daily wherever you get your podcasts.

[00:04:54] So Anil, I've been thinking about how not long after we invented digital computers, we started referring to our human brains as computers.

[00:05:03] Obviously, there is a lot more to it than that.

[00:05:06] But what is helpful and not helpful about describing our brains as computers?

[00:05:10] It's clearly very helpful.

[00:05:12] I mean, my title, my academic title is Professor of Computational Neuroscience.

[00:05:17] So I'd be rather hypocritical to say that it was not a useful way of thinking to some extent.

[00:05:23] And there's a very lively debate, mainly in philosophy rather than neuroscience or in tech,

[00:05:30] about whether brains actually do computation as well as other things.

[00:05:36] In fact, the metaphor of the brain as a computer has clearly been very, very helpful.

[00:05:42] If you just look inside a brain, you find all these neurons and chemicals and all kinds of complex stuff.

[00:05:47] And computation gives us the language to think about what brains are doing.

[00:05:52] That means you don't have to worry so much about all that.

[00:05:56] And of course, at the beginning of AI, there was this idea that intelligence might be a matter of computation.

[00:06:02] Alan Turing famously asked the questions about whether machines can think.

[00:06:07] And universal Turing machines were specified theoretically, which can do any computation.

[00:06:14] And the idea that, well, that might be what the brain is doing becomes very appealing.

[00:06:18] Also at the birth of AI, Walter Pitts and Warren McCulloch realized that neural networks,

[00:06:24] these simple abstractions of artificial neurons that are connected to each other,

[00:06:30] that underpin a lot of the modern AI we have, actually serve as universal Turing machines.

[00:06:36] So we have this temptation, this idea to think, yeah, the brain is a network of neurons.

[00:06:41] Networks of neurons can be universal Turing machines.

[00:06:44] And these are very powerful things.

[00:06:46] So maybe the brain is a computer.

[00:06:48] But I think we're also seeing the limits of that metaphor and all the ways in which actually brains might perform computations,

[00:06:58] but they may also do other things.

[00:07:00] And fundamentally, we always get into trouble too when we confuse the metaphor for the thing itself.

[00:07:05] I love that.

[00:07:06] And I think a big chunk of that is also we talk so much about sort of the supercomputing clusters

[00:07:12] and just how fast technology is moving.

[00:07:14] And we're almost, you know, losing some appreciation for the intelligence that's inside our craniums.

[00:07:20] And to put it very plainly, how much more complex is the brain today compared to even the most advanced AI systems?

[00:07:28] I mean, it's a totally different thing.

[00:07:31] I think we really do the brain a great disservice if we think of it purely in terms of sort of number of neurons.

[00:07:38] But even then, there are 86 billion neurons in the human brain and a thousand times more connections.

[00:07:43] It's incredibly complicated, even at that level.

[00:07:46] Also, the brain is so intricate.

[00:07:50] Like the connectivity in one area might be slightly different from the connectivity in another area.

[00:07:55] There are also neurotransmitters washing around.

[00:07:59] The brain changes.

[00:08:00] Every time a neuron fires, synaptic connectivities change a little bit.

[00:08:04] It's not a stable architecture.

[00:08:08] And then there are all the glial cells and all the supporting gubbins that we often don't even think about,

[00:08:15] but are turning out to actually be significantly involved in the brain's function.

[00:08:20] There was a recent paper in Science, I think, that had this gargantuan, impressive effort

[00:08:27] to unpack in as much detail as possible one cubic millimeter of brain tissue in the human cortex.

[00:08:35] In this one cubic millimeter, you've got 150 million connections, nearly 60,000 cells.

[00:08:41] And to store all that data in a standard computer was just an enormous amount to characterize.

[00:08:47] And even this is just a, you know, it's not everything, right?

[00:08:50] This is just a very detailed model.

[00:08:52] The brain is very complex.

[00:08:54] Very complex.

[00:08:55] That is quite amazing.

[00:08:56] What's also interesting about the sheer complexity in the brain is the brain doesn't sit in a vat, right?

[00:09:01] At least not usually.

[00:09:02] And of course, the brain works in concert with the rest of the body.

[00:09:05] Does that aspect of being embodied give humans any advantages over AI systems?

[00:09:11] I think it depends what you want the system to do.

[00:09:14] You're absolutely right.

[00:09:15] Brains didn't evolve in isolation.

[00:09:18] They evolved in response to certain selection pressures.

[00:09:21] What were those selection pressures?

[00:09:22] They weren't sort of right computer programs or right poetry or solve complex problems.

[00:09:28] Fundamentally, brains are in the business of keeping the body alive and later on moving the body around.

[00:09:35] So control of action, things like that.

[00:09:37] And those imperatives, fundamental to me, to understanding what kinds of things brains are,

[00:09:45] they are part of the body.

[00:09:46] They're not these kind of meat computer that moves this body around from one meter to another.

[00:09:53] Chemicals in the body affect what's happening in the brain.

[00:09:57] The brain communicates with the gut, even with the microbiome.

[00:10:01] We're seeing all these kinds of effects that transpire within the body.

[00:10:06] And then, of course, the body is embedded in a world.

[00:10:09] And there's always this feedback from the world.

[00:10:12] And understanding these nested loops of how the brain is embedded within a body and the body is embedded within a world,

[00:10:20] I think that's a very different kind of thing than the abstract, disembodied ideal of computation

[00:10:28] that drives a lot of our current AI.

[00:10:32] And, of course, it's also represented in a lot of science fiction.

[00:10:34] And then we have things like how, which are 2001, which, OK, there's a body as the spaceship,

[00:10:40] but it's a kind of disembodied intelligence in many ways.

[00:10:44] So then how important is it that we're embodied to have consciousness and intelligence?

[00:10:48] And we'll get to the definitions in a bit because I'm curious what happens when you embody an AI.

[00:10:54] And I'm, of course, thinking of all the humanoid robot demos that we've seen lately,

[00:10:57] where it seems to be this crude representation of kind of what we do.

[00:11:01] Like we've got sensor systems that perceive the world and we build a map of it

[00:11:05] and then we can figure out how to take action in it.

[00:11:08] This is a very, this is a fascinating question.

[00:11:10] I mean, so far, you know, the AI systems that we have,

[00:11:13] the ones we tend to hear about mainly anyway, language models and generative AI,

[00:11:18] they tend to have been trained and then deployed in a very disembodied way.

[00:11:23] But this is, this is changing.

[00:11:24] And robotics is improving too.

[00:11:26] It's lagging behind a little bit as it always does, but it is improving.

[00:11:29] And there are fascinating questions about what difference that makes.

[00:11:34] One possibility that strikes me as plausible is that in bodying an AI system so that you train it

[00:11:42] in terms of physical interactions, don't just drop a pre-trained model into a robot,

[00:11:47] but everything is trained in an embodied way, might give us grounds to say that AI systems

[00:11:53] actually understand what they say, if it's a language model, for instance, or understand what

[00:12:00] they do.

[00:12:01] Because these abstract symbols, words that we use in language, and there's a good argument

[00:12:06] that ultimately their meaning is grounded in physical interactions with the world.

[00:12:11] But does this mean that AI systems not only are intelligent and possibly understand it,

[00:12:16] but also have conscious experiences?

[00:12:18] That's a separate question.

[00:12:19] And I think there's many other things that might be necessary for us to think seriously

[00:12:28] about the possibility of AI being conscious.

[00:12:30] I think that brings me to the logical next question, which is,

[00:12:35] what is the difference between intelligence and consciousness?

[00:12:38] Perhaps let's start with intelligence.

[00:12:40] Both intelligence and consciousness are tricky to define,

[00:12:43] but most definitions immediately point to the fact that they're different.

[00:12:46] And if we think about a broad definition of intelligence, it's something like doing the

[00:12:52] right thing at the right time.

[00:12:54] A slightly more sophisticated definition might be the ability to solve problems flexibly.

[00:13:00] And whether it's solving a Rubik's Cube or a complex problem scientifically or navigating

[00:13:06] a social situation adeptly, I mean, these are all aspects of doing the right thing at the

[00:13:11] right time.

[00:13:12] And importantly, intelligence is something you can define in terms of function, in terms

[00:13:17] of what a system does, what its behavior ultimately is.

[00:13:21] So there's no deep philosophical challenge for machines to become intelligent in some way.

[00:13:28] I mean, there may be obstacles that prevent machines from becoming intelligent in this sort

[00:13:33] of general AI way, which is the way that humans are intelligent.

[00:13:37] But intelligence fundamentally is a property of systems.

[00:13:41] Now, consciousness is different.

[00:13:44] Consciousness, again, is very hard to define in a way that's going to get everyone signed up to.

[00:13:50] But fundamentally, consciousness is not about doing things.

[00:13:55] It's about experience.

[00:13:56] It's the difference between being awake and aware and the profound loss of consciousness

[00:14:02] in general anesthesia.

[00:14:03] And when you open your eyes, your brain is not merely responding to signals that come

[00:14:09] into the retina.

[00:14:10] There's an experience of color and shape and shade that characterizes what's going on.

[00:14:17] A world appears and a self within it.

[00:14:20] Thomas Nagel, I think, has the nicest philosophical definition, which is that for a conscious organism,

[00:14:25] there is something it is like to be that organism.

[00:14:29] It feels like something to be me.

[00:14:31] It feels like something to be you.

[00:14:32] Now, you can finesse these distinctions as much as you want, these definitions.

[00:14:38] But I think it's already clear.

[00:14:40] They are different things.

[00:14:41] They come together in us humans.

[00:14:43] We know we're conscious and we think we're intelligent.

[00:14:46] So we tend to put the two together.

[00:14:48] But just because they come together in us doesn't mean they necessarily go together in general.

[00:14:54] As you describe consciousness and this sort of subjective experience, a term that keeps

[00:14:58] getting thrown around in AI circles now is like qualia, right?

[00:15:01] This notion of subjective conscious experiences and figuring out if large language models can

[00:15:06] actually have this.

[00:15:08] Certainly, they're good at making it seem like they do, especially the jailbroken models.

[00:15:13] But it also takes me back to something else that you've talked about, which is our perception of

[00:15:17] reality is sort of this controlled hallucination that we don't fully perceive reality in this

[00:15:23] completely objective sense.

[00:15:24] I don't know if that's the best characterization.

[00:15:26] But I'm trying to connect the dots there where it seems to be like even our experience of reality

[00:15:32] is kind of hard to grok and fully explain.

[00:15:35] And so I wonder, doesn't that point to us not being able to create a very clear definition

[00:15:41] to measure that in a synthetic system?

[00:15:43] Yeah, I think you can go even further, actually.

[00:15:45] I think there's very little consensus on, well, there's no consensus on what would be the necessary

[00:15:51] and sufficient conditions for something to have subjective experience, to have qualia in this

[00:15:58] sense.

[00:15:58] When you and I open our eyes, we have a visual experience.

[00:16:03] It's the redness of red, the greenness of green.

[00:16:06] This is the kind of thing that philosophers call qualia.

[00:16:09] And there's a lot of argument about whether this is actually a meaningful concept or it's

[00:16:13] just something that we think is profound and it actually is just a wrong way of looking

[00:16:18] at the problem.

[00:16:18] But for me, there is a there-there.

[00:16:22] When we open our eyes, there is visual experience.

[00:16:24] However, we label it as qualia or something else.

[00:16:28] But for a camera on my iPhone, well, no.

[00:16:32] We don't think there's any experiencing going on.

[00:16:36] So what is the difference that makes a difference?

[00:16:39] And could it be that some kind of AI that's a glorified version of my camera on the phone

[00:16:45] would instantiate the sufficient condition so that it not only responded to visual signals,

[00:16:51] but had subjective experience too?

[00:16:55] I think that's the challenge we need to face because, as you say, AI systems, especially

[00:17:00] things like language models, can be very persuasive about having conscious experience.

[00:17:07] And again, especially the ones where you ask them to whisper and get around the guardrails

[00:17:13] in one way or another.

[00:17:15] They can really seduce our biases.

[00:17:16] And so we can't just rely on what a language model says.

[00:17:23] If a language model says, yes, of course, I have a conscious visual experience, that's

[00:17:28] not great evidence for whether it's there or not.

[00:17:31] And so we need to think, I think, a little more deeply about what it would take to ascribe

[00:17:36] conscious experience to the system that we create out of a completely different material.

[00:17:43] And material is an interesting point you're bringing up, sort of the substrate from which

[00:17:47] intelligence and perhaps consciousness can emerge.

[00:17:50] Because what you are saying is that I think it seems clear that we could have a superhuman

[00:17:56] intelligence level AI system that isn't necessarily conscious.

[00:18:00] But I do wonder when people make arguments like, hey, well, if we just keep throwing more

[00:18:04] data and compute at this thing and it keeps getting more and more intelligent, consciousness

[00:18:08] will be this emergent property of this system.

[00:18:10] And it almost has this like techno-religious kind of fervor to it.

[00:18:14] Why do you think consciousness might be uniquely biological?

[00:18:18] Why does the nature of the substrate matter?

[00:18:20] I don't know that it matters, but I think it's a possibility worth taking seriously.

[00:18:27] In a sense, the opposite claim is equally odd.

[00:18:30] Why should consciousness be a property of a completely different kind of material?

[00:18:35] Why would computation be sufficient for consciousness?

[00:18:40] After all, for many things, the material matters.

[00:18:45] If we're talking about a rainstorm, you need actual water for anything to get wet.

[00:18:49] If you have a computer simulation of a weather system, it doesn't get wet or windy inside

[00:18:55] that computer simulation.

[00:18:56] It's only ever a simulation.

[00:18:58] And the way you set it up is also very informative because there has been this implicit assumption,

[00:19:05] at least in some quarters, that indeed if you just throw more compute and AI gets smarter

[00:19:11] in ways which can be very, very impressive and very, very sometimes unexpected too, that

[00:19:17] at some point consciousness just arrives, comes along for the ride and the inner lights come

[00:19:22] on and you have something that is also experiencing as well as being smart.

[00:19:28] I think that's a reflection more of our psychological biases than it is grounds for having credence

[00:19:36] in synthetic consciousness.

[00:19:39] Because why should consciousness just happen at a particular level of intelligence?

[00:19:44] I mean, you could make an argument that some forms of intelligence might require consciousness.

[00:19:50] And those may be the kinds of intelligence that we humans have.

[00:19:57] But that's a bit of a strange argument because there are plenty of other species out there

[00:20:01] that don't have human-like intelligence that are very likely conscious.

[00:20:05] And there may be more ways to achieve intelligence than through what evolution settled on for human

[00:20:13] beings, which is having brains that are also capable of consciousness.

[00:20:17] So the question for me, the fundamental question is, is computation sufficient for consciousness?

[00:20:24] If we try to design in the functional architecture of the brain as it is and run it in a computer,

[00:20:31] would that be enough for consciousness?

[00:20:33] Or do we need something much more brain-like at the level of being made,

[00:20:39] of carbon, of having neurons, of having neurotransmitters washing around,

[00:20:43] of being really grounded in our living flesh and blood matter?

[00:20:49] And I don't think there's, well, there's just not a knockdown argument for or against either

[00:20:55] of these positions.

[00:20:57] But there's, to me, good reasons to think that computation is likely not enough.

[00:21:03] And there are at least some good reasons to think that the stuff we're made of really does matter.

[00:21:07] Given all of this, you do believe that it's unlikely that AI will ever achieve consciousness.

[00:21:13] Why is that?

[00:21:14] I think it's unlikely, but I have to say it's not impossible.

[00:21:17] And the first reason it's not impossible is that I may very well be wrong.

[00:21:22] And if I'm wrong and computation is sufficient for consciousness, well, then it's going to be a lot

[00:21:28] easier than I think.

[00:21:30] But even if I'm right about that, then as AI is evolving and as our technology evolves,

[00:21:38] we also have these technologies that are becoming more brain-like in various ways.

[00:21:44] We have these whole areas of neuromorphic engineering or neuromorphic computing,

[00:21:49] where we're building systems which are just sticking closer to the properties of real brains.

[00:21:58] And on the other side, we also have things like cerebral organoids, which are made out of brain

[00:22:03] cells. They're little mini brain type things grown in the dish. They're derived from human

[00:22:09] stem cells and they differentiate into neurons, which clump together and show organized patterns

[00:22:14] of activity. Now they don't do anything very interesting yet. So it's the opposite situation

[00:22:20] to a language model. A language model really seduces our psychological biases because it speaks

[00:22:26] to us. But a clump of neurons in a dish just doesn't because it doesn't do anything yet.

[00:22:32] Now, for me, the possibility of artificial consciousness there is much higher because

[00:22:36] we're made out of the same material. To the specific question, why should that matter? Why does

[00:22:42] the matter matter? It comes back to this idea about what kinds of things brains are and the fact that

[00:22:49] they're deeply embodied and embedded systems.

[00:22:51] So brains fundamentally, in my view, evolved to control and regulate the body, to keep the body alive.

[00:22:58] And fundamentally, this imperative goes right down, even into individual cells. Individual cells

[00:23:07] are continually regenerating their own conditions for survival. They don't just take an input and

[00:23:13] transform it into an output. And in doing this, I think there's pretty much a direct line from the

[00:23:20] metabolic processes that are fundamentally dependent on particular kinds of matter, flows of energy,

[00:23:27] transformations of carbon into energy, things like that, all the way up to these high level descriptions

[00:23:33] of the brain making a perceptual inference, or as we said earlier, a controlled hallucination of best

[00:23:39] guess about the way the world is. So if there is this through line from things that are alive and why we call them alive,

[00:23:49] all the way up to the neural circuitry that seems to be involved in visual perception or conscious experience

[00:23:55] generally, then I think there's some reason to think that consciousness is a property of living systems.

[00:24:02] As you were answering that, in my head, I have this visualization, maybe the future of this conscious AI system

[00:24:08] that we finally create isn't going to be a bunch of Jensen's NVIDIA GPUs and some data center, but perhaps this

[00:24:14] like giga brain that we build out of the very things that our brain is made out of. That's one hell of a visual,

[00:24:20] I got to say.

[00:24:21] Yeah, I think that's, and that's a possible future, right? Because we're already on that track with

[00:24:26] with neurotechnologies and hybrid technologies as well. And people can plug organoids into rack servers,

[00:24:34] people are beginning to do this already to sort of leverage the dynamical repertoire that these things have.

[00:24:42] And nobody knows how biological a system needs to be in order to move the needle on the possibility

[00:24:49] for consciousness happening. It may be not at all, or it may be a great deal indeed.

[00:25:07] So I have to ask the question, can artificial neural networks then also teach us something about

[00:25:13] biological neural networks? And the reason I asked this, I was reading the Anthropic CEO's

[00:25:18] rather extended blog, and he brought up this example where basically like a computational

[00:25:23] mechanism was discovered by AI interpretability researchers in these AI systems that was rediscovered

[00:25:29] in the brains of mice. And I was just asking my question, wait a second. So like,

[00:25:34] like an artificial system, like a very simplified simulation is still telling us something about the

[00:25:40] organic, you know, more complex representation. What's your thoughts on that? And do you think

[00:25:45] this trend will continue?

[00:25:46] Oh, absolutely. I think this is for me, the line of research that was certainly the line that I'm

[00:25:53] following. The use of computers and in general, in AI in particular, they're incredible tool.

[00:26:00] They're incredible general purpose tools for understanding things. And, you know, even in

[00:26:04] my own research, this is what we do. I mean, we'll build computational models of what we think is

[00:26:10] going on in the brain. And we'll see what these models are capable of doing. And we'll also see

[00:26:16] what predictions they might make about real brains that we might then go and test in experiments.

[00:26:22] I have to imagine the advances in technology, both on the sensing and the computation side is making

[00:26:27] a huge difference. And I'd love to hear some examples.

[00:26:30] There are examples in many different levels. So for instance, there are algorithms involved in

[00:26:35] generative AI that might really map onto things that brains do. So one level, it's about discovering

[00:26:44] what the functional architecture of the brain is through developing these new kinds of algorithms.

[00:26:49] But then there are other levels too, in that there's the levels in which we might use AI systems as tools

[00:26:57] for modeling or understanding some higher level aspects of the brain. So for instance,

[00:27:03] we use some generative AI methods to simulate different kinds of perceptual hallucination.

[00:27:09] So the visual hallucinations that people have in different conditions like in psychosis or in

[00:27:14] Parkinson's disease or after psychedelics. And this goes back to some early algorithms by Google in

[00:27:22] their deep dream when they turned bowls of pasta into these weird images with dogheads sprouting

[00:27:28] everywhere. But we can use those in a more serious way to get a handle on what's happening in the brain

[00:27:34] when people experience hallucinations. And then right at the other end, and I admit this is something that

[00:27:40] for me anyway, is still uncharted territory and something I'm really interested to explore

[00:27:46] is when we actually leverage the tool set that AI is delivering now, you know, the language models,

[00:27:53] the virtual agents. And I was reading a paper just the other day about a whole virtual lab that was

[00:27:59] discovering new compounds to bind to the COVID virus, virus particles. And this virtual lab was

[00:28:09] basically doing everything from searching literature to generating the hypothesis to critiquing

[00:28:15] experimental designs and proposing new experimental designs and so on. So I think there's a lot of

[00:28:20] utility in AI for accelerating the process of scientific discovery.

[00:28:26] I think AlphaFold is such a great example of that, right? Like what took like a PhD,

[00:28:32] you know, the entirety of their PhD to figure out a couple of molecules. We've mapped out a huge,

[00:28:38] huge opportunity space and kind of just put it out there.

[00:28:41] I mean, that is such a beautiful example because also it just exemplifies the way in which I think

[00:28:48] it's productive for us to relate to these kinds of systems because AlphaFold intuitively seems like

[00:28:54] a tool, right? We treat it, we use it as a tool or rather the biologists do to just rapidly accelerate

[00:29:01] the hypotheses they can make at the level of protein binding. We never think of AlphaFold as another

[00:29:08] conscious scientist. It's not, it doesn't, it doesn't seduce our intuitions in the same way that

[00:29:14] language models do. So I don't think there's anything quite comparable to AlphaFold, you know,

[00:29:22] in the neuroscience domain yet. And I'm trying to think what one, what, what the equivalent problem

[00:29:29] would be, you know, what one thing might be, and this is, this is very speculative. Maybe somebody's

[00:29:35] working on this already, but you know, one of the big unknowns in the brain is, is really how it's

[00:29:41] wired up. There was, uh, you know, another recent paper looking at the full wiring diagram of the

[00:29:47] brain of a fruit fly. And this is an incredible resource already. It was computationally incredibly

[00:29:53] difficult to, to put this together from the little bits of data that you might get in individual

[00:29:59] experiments. So there could well be a role for, for AI in helping amass large amounts of data

[00:30:04] to give us a more comprehensive picture of what kind of thing a brain is. Um, and there may be

[00:30:10] many other creative ideas out there too. Uh, but all of them I think would treat the AI

[00:30:17] as in the, in, I think the most productive way as a kind of tool.

[00:30:21] I agree. There's a, there's a lot of, you know, inclination to, I call it the co-pilot versus

[00:30:26] captain question. A lot of people are like, yeah, this is like my personalized Jarvis and I'm going to

[00:30:31] be like Tony Stark in the lab and just like, you know, doing what I need to do. And it just like

[00:30:35] preempts my needs. And it's cool that it's not constrained sort of by wall clock time, right? That

[00:30:40] you can just throw more compute at it and they can move faster. But fundamentally to me, it feels like

[00:30:44] humans are still doing the orchestration. Um, what do you think are the risks of going the other route

[00:30:50] where we start feeling like these systems should be the captain and let's build the grand AGI system

[00:30:55] and ask it what to do and then let's do it blindly. Yeah, I think, I mean, there's, there's, of course,

[00:31:01] there's a huge amount of uncertainty. Maybe it's not a, not a terrible idea in some ways, but it does

[00:31:06] strike me as something that is certainly not guaranteed to turn out very well. And human intuition

[00:31:12] still seems very important in interpreting the suggestions that might come from, from AI or

[00:31:18] just what AI will deliver in whatever context, having a human in the loop still seems to be very,

[00:31:24] very important. But there are some larger risks here that to the extent we do this, then I think

[00:31:30] we are moving back towards imbuing artificial intelligence with properties that it may not in fact

[00:31:40] have, you know, things like, oh, it really does understand what it's doing or, or it may be

[00:31:46] indeed be conscious of, of, of what it's doing as well. I think if we miss attribute qualities like

[00:31:53] this to AI, that can be pretty dangerous because we may fail to predict what it will do. Um, another

[00:32:03] concept from Daniel Dennett is something called the intentional stance. And it's a beautiful idea about

[00:32:08] how we interpret the behavior of other people is because I attribute beliefs and knowledge and goals

[00:32:15] to, to you or to whoever I'm interacting with. And that helps me predict what they're going to do.

[00:32:20] Now, if we, if we do this with AI systems, and this is what language models in particular encourages

[00:32:25] to do, then we may get it right some of the time, but we may get it wrong some of the time too,

[00:32:31] if the systems don't actually have these beliefs, desires, goals, and so on. And that can be, that can

[00:32:38] be quite problematic. There's the other side to all of this too, where, you know, technology is also

[00:32:43] advancing to a degree where, um, we can kind of coarsely figure out what's going on in people's

[00:32:49] minds. And so earlier in the season, we had Nita Farahani on, and she touched on the concept of

[00:32:54] cognitive liberty. And we were basically nerding out over how we're basically putting all these like

[00:33:01] ourselves. And yes, right now they can coarsely read our brains. And what was even trippier to

[00:33:06] learn about is manipulating our dreams with targeted dream incubation. What keeps you up at

[00:33:12] night when you think about sort of the ethical considerations from AI kind of making our minds

[00:33:18] more of an open box than they have been in the past? One of the things that I think about, and I was

[00:33:23] recently writing a paper with a philosopher, Emma Gordon, about brain computer interfaces as well,

[00:33:28] is, is really, why is the skull, this, this boundary that we think of as particularly significant

[00:33:34] here? I mean, we've already given our data privacy away in so many ways. And that's true,

[00:33:39] not a good thing, right? But, but in many ways, at least for people who've been around for a while,

[00:33:45] the cat's already out of the bag, but the idea of getting inside the skull does seem to be

[00:33:52] significant, partly because there's no other boundary that's left. And while we're very used to the

[00:33:58] is the importance of things like preserving freedom of speech, then there isn't really the

[00:34:04] same degree of attention paid to something like freedom of thought. Right.

[00:34:07] So we just not used to what kinds of guardrails and moral guidelines we might need in this case.

[00:34:15] And then there are also, I think, some more subtle worries, certainly in this space of brain computer

[00:34:20] interfaces, because let's imagine a situation where we each have, or brain computer interfaces are widely

[00:34:26] used. A lot of brain data is extracted, it's used to train models, which are then used to underpin the

[00:34:34] utility of brain computer interfaces so that they can predict what someone wants to say or do on the

[00:34:40] basis of brain activity. Now there are some extraordinarily powerful and compelling use cases

[00:34:45] for this kind of thing in medicine, in treatment for people with brain damage or paralysis or blinds.

[00:34:52] But if we generalize that to enhancement of everybody, and we try to think, okay, these things are not just

[00:34:59] a soul specific clinical problems, but they become part of our society more deeply, then there's a

[00:35:06] potential that there's a kind of enforced homogeneity. Yeah, we might have to learn to think a

[00:35:11] particular way in order to get the brain computer interface to work. And that may be a completely

[00:35:16] unintended consequence. But it strikes me as a worrisome consequence as well. There may also be,

[00:35:23] you know, just kinds of social inequity that start to happen to about, okay, people with access to

[00:35:29] these systems can do more, or will be allowed to train them so they can think in their own distinctive

[00:35:33] way and not have to think in, you know, in the way that the mass market BCIs require. So I think

[00:35:39] there's a lot of, there's a lot of sort of feedback cycles that can start to unfold in this case. But

[00:35:46] fundamentally, it's that really, there's nothing more to privacy once you go inside the skull. And

[00:35:53] then there's a stimulation thing as well. Brain computer interfaces can be bi-directional. And if

[00:36:00] they're bi-directional, and you start imprinting thoughts, goals, intentions, then we're definitely

[00:36:09] in a very ethically troubling situation. That last bit to me is the stuff that keeps me up. It's like,

[00:36:15] it's like giving a bunch of companies read-write access to your mind, right? And to your brain. And

[00:36:21] in a sense, the point you brought up about sort of, you know, homogeneity, sort of like lack of

[00:36:27] intellectual diversity, we're already kind of seeing that where people are using LLMs and it's

[00:36:31] all kind of like the same milk toast prose. And, you know, people are almost losing the ability to

[00:36:38] write and think. And yeah, I think there's something kind of disconcerting about that.

[00:36:45] Yeah. I mean, there might be a more optimistic view of this too, that the sort of

[00:36:49] milk toast homogeneity of large language model output may cause us to really value human contributions,

[00:36:56] more. You know, just as in other situations, there's a kind of, there's a value attached to

[00:37:01] the handmade, the bespoke. And we may end up living in a situation where we just view these

[00:37:09] two kinds of language quite differently. And just as someone who grows up in a bilingual household,

[00:37:15] you know, will naturally learn to speak two different languages. Future generations might

[00:37:19] become accustomed to, okay, well that's kind of large language model language. And this is human

[00:37:24] language. And they just innately feel very different, even though they're using the same words.

[00:37:30] Oh yeah. Kind of like just forming code switching and, you know, different contexts,

[00:37:33] how exactly to behave. I think that's a, it's a valid point. Perhaps I have a slightly more jaded

[00:37:38] take on this. Cause I'm like, yeah, people are going to want the whole foods experience,

[00:37:43] but a vast majority of people are like, give me the free ad funded mountain do straight to the vein.

[00:37:47] And I deeply, deeply worry about that. So let's leave the lab for a second here. What are the kinds

[00:37:53] of AI tools that you yourself are using outside of the context of the lab?

[00:37:57] I'm pretty a light user of, of AI tools, at least the ones that I know about, because of

[00:38:02] course one of the thing is AI is hidden beneath the surface of many of the things we use. And

[00:38:07] every time I use Google maps, it says, you know, there's machine learning or AI happening there.

[00:38:11] I do use language models increasingly sort of as verbal sparring partners, rather than as sources of text

[00:38:19] that I will then edit or use directly. And it's a kind of as glorified search engines in that sense.

[00:38:27] And yeah, I find them more and more useful, but I still don't trust them. I think it's a case of

[00:38:34] using them to help us, to help humans think more clearly rather than to outsource the business of

[00:38:40] thinking itself.

[00:38:41] So have you ever, whilst interacting with all these large language models, felt yourself forming

[00:38:48] a connection with these systems? Or are you able to keep that separation and distance? Like almost

[00:38:54] like you're forgetting it's a tool and kind of more like a colleague. Does it ever feel like that?

[00:38:58] It doesn't. You know, this is another of the things that keeps me up at night, back to that, that

[00:39:03] question. Um, because there's something so seductive about the way we respond to language that even if

[00:39:11] at one level, we can be very, very skeptical that there's anything other than just sort of statistical

[00:39:18] machination happening, the feeling that there's a mind that understands and might be conscious is

[00:39:25] extremely powerful. And I'm thinking, you know, one way of thinking about this is that there are plenty

[00:39:31] of cases where knowledge does not change experience. So for example, lots of visual illusions. Um,

[00:39:40] there's a famous visual illusion called the Muller-Layer illusion, which is a visual illusion where

[00:39:45] two lines, um, look different lengths because of the way the arrows point at the end, but

[00:39:51] you measure them, they're exactly the same length. And the thing is, even if you know this, even if you

[00:39:57] understand what's happening in the visual system that gives rise to this illusion,

[00:40:00] they will always look, you know, the way they do.

[00:40:04] There's no like firmware, there's no firmware fix for our brains to fix that.

[00:40:08] That's right. And so the worry for me is that there will be similarly cognitively impenetrable

[00:40:15] illusions of artificial consciousness. That if we're dealing with sufficiently fluent language models,

[00:40:21] especially if they get animated in deep flakes or even embodied in humanoid robots,

[00:40:27] that we won't be able to update our own wetware sufficiently in order to not feel that they

[00:40:36] are conscious. We will just be compelled to have those kinds of feelings. And that is a very

[00:40:44] problematic state to land in too, because if we are unable to avoid attributing, let's say conscious

[00:40:50] states to assist them. And then again, we're going to be in the business of attributing with qualities

[00:40:56] it doesn't have and mispredicting what it's going to do and leaving ourselves more open to coercion,

[00:41:02] and more vulnerable to manipulation. Because if we think a system really understands us and cares about

[00:41:07] us, but it doesn't, it's actually just trying to sell us Oreos or something, then that's a problem.

[00:41:13] And I think that the most pernicious problem here is something that goes right back to Immanuel Kant and

[00:41:19] probably before, which is the problem of brutalizing our own minds. Because here, if we are interacting

[00:41:27] with an artificial system that we can't help but feel is conscious, we have two options broadly. We

[00:41:34] can either be nice to it anyway and care about it and bring it within our circle of moral concern.

[00:41:40] And that's okay. But it means that, you know, we will waste some of our moral capital on things that

[00:41:45] don't need it and potentially care less about other things because we humans have this in-group,

[00:41:50] out-group dynamic. If you're in, you're in. If you're out, you're out. So we might either do that

[00:41:56] or we learn to not care about these things and sort of treat them in the same way that we might

[00:42:02] treat a toaster or a radio. And that can be very bad for us psychologically because if we treat

[00:42:09] things badly, but we still feel they are conscious, that's the point that Kant made. That's what

[00:42:14] brutalizes our minds. It's why we don't poke the eyes out of teddy bears or pull the limbs off dolls.

[00:42:21] You know, the science fiction film and series Westworld dealt with this beautifully. You know,

[00:42:26] how dangerous it is for us to take this perspective.

[00:42:31] So this keeps me up at night because there's no good option here. We need to think very carefully,

[00:42:37] not only about the possibility of designing actually conscious machines, which even if it is

[00:42:44] unlikely, if it happened, would be very ethically problematic because of course, if something actually

[00:42:49] is conscious, it's a moral subject and we would need to be very careful about how we treat it. But even

[00:42:56] building systems that give the strong appearance of being conscious is also problematic for different

[00:43:02] reasons. And this scenario is basically already with us. It will be very soon unless we think very

[00:43:09] carefully about how we design these systems and design against giving that impression in some way.

[00:43:15] I think you very beautifully paint this picture of why it's problematic on both ends, right? Like

[00:43:19] treating it like the Rick and Morty, this robot wakes up, what is your purpose? Your purpose is to

[00:43:24] put butter on my toast. That is your purpose. Just get back to please putting butter on my toast.

[00:43:28] And it has this existential crisis. And I think on the other end, the Westworld example is very valid

[00:43:34] too, where you have things that are indistinguishable from humans and we go act out all these sort of

[00:43:39] lower urges or whatever the right way to put that. And we suddenly start bringing that

[00:43:45] sort of behavior to interactions with actual humans. But the real question I come at is where

[00:43:49] you end it, which is from a user experience standpoint, right? A lot of people think that it is

[00:43:55] important to have these systems be as human-like as possible and meet the users sort of where they are.

[00:44:01] Do you want to talk about why we need to be more nuanced? And do you have any ideas for what that sort of

[00:44:08] would be a better way to build these systems? Because it seems like either extreme kind of sucks.

[00:44:14] I think this is super interesting. And in fact, I think talking to you just now is helping

[00:44:19] give focus to this as a serious design challenge. And I'm not sure it's one that's been well addressed

[00:44:25] so far. And because of course, yes, there is a good reason to build systems with which we can

[00:44:32] interact very fluently. It can also be very empowering. If we can have a machine generate

[00:44:38] code by talking to it about what we want a program to do, that's hugely empowering for many people,

[00:44:44] so long as it does the thing that it's supposed to do and not something else. But is there a way of

[00:44:51] having the benefits of that, designing systems so that we can preserve the kind of fluent interaction

[00:44:58] that natural language gives us, but in a way that still at least pushes back to some extent on the

[00:45:07] psychological biases that then lead us to make all these further attributions of consciousness,

[00:45:15] of understanding, of caring, of emotion, and all of these things. I don't know what the solution is,

[00:45:21] but I think it's a really important problem. One simple solution would be, okay, these things just

[00:45:28] have to watermark themselves to say, I am not conscious, I don't have feelings. And of course,

[00:45:33] language models do that until you play around with them, press them a little bit. But that may not be

[00:45:40] enough. There may have to be other ways where we design interfaces which, through practice or through

[00:45:48] education or through some other manipulation are shown, and this is really a question as much for

[00:45:55] psychology as it is for technology. What kinds of things preserve fluid interaction but do not

[00:46:04] succumb to our psychological biases and what properties we attribute? I would love to see

[00:46:10] focus on that problem because that would show the line we need to walk.

[00:46:16] And you're right that there aren't any solutions. Do you think we can build those antibodies?

[00:46:22] I think we have to try. I mean, that also brings up another point, which is again, very contentious

[00:46:28] in the tech sphere, which is what should we do about regulation? What kinds of systems should people

[00:46:34] just put out there? And what I come back to in that conversation is always the fact that in other

[00:46:42] domains of invention and technology, we're very cautious. We don't put a new plane in the sky

[00:46:49] without being fairly sure it's not going to fall out. We don't release a new drug on the market unless

[00:46:54] we can be very sure it's not going to have unintended side effects or consequences. And there does seem to

[00:47:01] be an increasing recognition that AI technologies are in the same ballpark. And doesn't mean that we want to

[00:47:11] stifle innovation, of course, but we can help shape and guide innovation. I think there's again a sweet spot to be found

[00:47:19] there. And then on the other side, the education, one of the challenges there, of course, is that things are moving so fast,

[00:47:24] that it's very hard to keep up. But it's important to try. One thing that strikes me here is the very

[00:47:32] term artificial intelligence is part of the problem. It brings with it so much baggage that there's some

[00:47:39] kind of magic and it's like, you know, science fiction mind, whether it's, you know, Javis or whether

[00:47:43] it's Skynet and Hell from 2001 or whatever it is, whatever your favorite conscious intelligent robot is,

[00:47:50] that's what we think of. And, you know, artificial intelligence has this brand quality, which I

[00:47:55] think is a little bit unhelpful. It may be, you've been incredibly successful in raising large amounts

[00:47:59] of venture capital, but it's, you know, it's not a particularly helpful description of what the systems

[00:48:04] themselves are doing. Of course, most people working in this, at least they used to at any rate,

[00:48:10] talk about machine learning rather than artificial intelligence. And then another level of

[00:48:14] description, you can just say, well, these things are basically just applied statistics.

[00:48:18] You start describing something as applied statistics, you know, it, it, it's even that

[00:48:26] is educationally valuable because it highlights how much we, when you load onto these systems

[00:48:33] by the words we use. One other very simple example here, which I think the, you know, the

[00:48:40] horse has already bolted here too, but it's always annoyed me how people describe language models

[00:48:46] as hallucinating when they make stuff up.

[00:48:49] Yeah. It's giving them too much credit.

[00:48:51] David It's giving them way too much credit. And it's,

[00:48:53] it's doing something more specific than that. It's cultivating the implicit idea that indeed

[00:49:00] language models experience things because that's what a hallucination is. A hallucination,

[00:49:05] if you apply it to a human being, it means, well, they're,

[00:49:08] they're having a perceptual experience of something that that's not there.

[00:49:12] David And the fact that that linguistic term caught on so quickly, I think is itself telling

[00:49:18] because it just revealed like, okay, there's some implicit assumptions about what people

[00:49:23] think these things are doing, but it's also in a positive feedback. It's unhelpful because it leads

[00:49:28] us to again, project qualities in. If we're going to use a word, I wish they'd used confabulation

[00:49:34] because in human psychology, confabulation is what people do when they make stuff up without

[00:49:40] realizing they're making stuff up. And that to me is an awful lot closer to what language models

[00:49:45] do, but I, yeah, I don't think it's going to catch on now, but we should be careful about the

[00:49:49] language we use to describe these systems for exactly this reason.

[00:49:53] David What advice would you have for people that are listening to this

[00:49:57] so that they can take advantage of the tools at their disposal today and not get sucked into,

[00:50:03] into perhaps the, I don't want like the pseudoscience and the, you know, fake spirituality that kind of comes

[00:50:09] as a package deal with AI today.

[00:50:12] David Part of it is exactly recognizing these often implicit motivations that drive all these,

[00:50:18] these associations that lead us to think of these things as being more than they are or different

[00:50:23] than they are. And in the extreme, it gets, it gets pretty religious, right? I mean, there's,

[00:50:30] there's the idea of the singularity, which is a sort of the techno-optimist moment of rapture and

[00:50:36] the possibility of uploading to the cloud and living forever with the promise of immortality.

[00:50:43] David Part of it is very, it's textbook religion, right? And so that in itself, I think is useful

[00:50:50] to bear in mind that there's a, there's a larger cultural story behind this. It's not simply an

[00:50:55] objective description of where the technology is. And then cashing that out further, it is for me,

[00:51:03] anyway, it's just a matter of, of continually reminding myself of the differences between us

[00:51:10] and the technologies that we build to resist the temptation to anthropomorphize, you know,

[00:51:15] to project human-like qualities into things, to retain a slightly critical attitude to what's going on

[00:51:22] behind the interface, that it is not an alternative person. And this can be easier said than done, because

[00:51:30] as we were discussing before, one of the things that keeps me up at night are these cognitively

[00:51:35] impenetrable illusions of intelligence, consciousness, and so on that AI systems can,

[00:51:41] can bring to bear. But yeah, I think it's just at its most basic, reminding ourselves that if we think

[00:51:50] an AI system is a reflection of us, that it's something in our image, what we're probably doing is

[00:51:57] overestimating its capabilities and underestimating our own capabilities.

[00:52:02] I love that. That's punchy. And that brings me to the last question,

[00:52:06] which is, given our discussion thus far, it makes me very curious, what is your idea of the sort of

[00:52:12] ultimate final form of AI, if you will, that appeals to you as a neuroscientist? Like what excites you the

[00:52:19] most about the potential for this future, you know, where, you know, AI can serve human intelligence

[00:52:26] and consciousness?

[00:52:27] David Morgan This is a very,

[00:52:28] very good question. I mean, I think my vision, optimistic vision about this is not some sort of

[00:52:36] single super intelligence, like Deep Thought and Hitchhiker's Guide to the Galaxy or whatever it might be.

[00:52:41] David Morgan That's a single super intelligent entity. Maybe AI in the future is going to be a bit more

[00:52:48] like electricity or water. You know, it's a basic utility of the utility. It's a utility and it's used

[00:52:56] in many, many different ways, in many, many different contexts to do many, many different things. And it's

[00:53:04] in this world, we face the challenge of, of recognizing that there are some things which we once thought

[00:53:11] were distinctively or uniquely human, which aren't. And so there will be a social disruption to that.

[00:53:17] This happens, of course, in all technological revolutions. But the flip side of that is that

[00:53:24] the space that's opened for massive innovation, creativity, the ability to solve all sorts of

[00:53:30] problems. So I think it's not a single thing. It's, it's many things. One last thought on this.

[00:53:36] I've heard the idea many times that the distinctive thing about AI is that it could be humanity's last

[00:53:41] invention because AI systems can design, develop, improve themselves.

[00:53:47] David Morgan Oh, they'll invent everything else?

[00:53:48] David Morgan Or we, we lose the ability to have dominion over what they may end up being. And that's

[00:53:56] something that I'm still, you know, I'm still a little bit unsure how to think about that, whether

[00:54:00] that's a, that's a real difference or whether it's something that we can still need to be careful to,

[00:54:05] to manage. But my, yeah, my optimistic view of AI is as some kind of utility that's drawn on in many,

[00:54:12] many ways. David Morgan Yeah, that permeates everything versus the,

[00:54:15] the singular all-encompassing AI in the sky, which again, gets start sounding very religious.

[00:54:21] Anil, thank you so much for joining us.

[00:54:23] Anil Thanks for the conversation. I really enjoyed it.

[00:54:28] Wow. What a conversation. Anil Seth reminds us that the story of AI isn't just a tale of machines

[00:54:35] gaining power. It's a mirror reflecting our own biases, aspirations, and fears. We project so much

[00:54:43] of ourselves onto these tools. We anthropomorphize the algorithms. We give meaning to their outputs,

[00:54:50] as if they share the complexity of human experience. But as Anil said, we overestimate their capabilities,

[00:54:58] and underestimate our own. And that's something worth meditating on. For all the dazzling feats AI can

[00:55:05] pull off, it's still us, humans, who design, direct, and decide what these systems become.

[00:55:12] And it's our responsibility to tread carefully, not just for the sake of innovation, but for the future

[00:55:18] of what makes us human. There's another fascinating question here too. If a truly conscious AI system

[00:55:25] does emerge one day, will it even look like the systems we built so far? The unique, messy biology of

[00:55:32] the human brain, neurons, synapses, and glial cells doesn't just power intelligence. It creates the rich,

[00:55:39] subjective experience we call consciousness. Silicon and software might never be enough.

[00:55:46] Consciousness may demand a substrate that mirrors what we're made of, a construct that's alive,

[00:55:52] pulsing with the same kind of vitality that flows through us. And this is wild to imagine,

[00:55:58] a future where conscious AI doesn't emerge from humming server racks or massive data centers,

[00:56:05] but from living organic systems, gigabrains we create from the very building blocks of life itself.

[00:56:13] In trying to recreate the sparks of consciousness, we'd be stepping closer to understanding what makes

[00:56:19] it so mysterious and so uniquely human. For now though, that's all in the realm of speculation.

[00:56:25] What's not speculative though is this. The choices we make today, how we design, interact with, and

[00:56:33] regulate these systems will shape not just the future of AI, but the future of our own humanity.

[00:56:40] And as much as we might marvel at the power of these tools, it's our responsibility to stay grounded,

[00:56:46] to remind ourselves that these systems are reflections of us, not replacements for us.

[00:56:53] It's truly an incredible moment to be alive, and a terrifying one too. And as we grapple with the

[00:56:59] unknowns ahead of us, perhaps the best question we can keep asking is, what kind of future do we want

[00:57:06] to create, not just for AI, but for ourselves? The TED AI Show is a part of the TED Audio Collective,

[00:57:18] and is produced by TED with Cosmic Standard. Our producers are Dominic Gerard and Alex Higgins.

[00:57:24] Our editor is Banban Chen. Our showrunner is Ivana Tucker, and our engineer is Asia Pilar Simpson.

[00:57:31] Our technical director is Jacob Winnick, and our executive producer is Eliza Smith.

[00:57:36] Our researcher and fact checker is Jennifer Nam. And I'm your host, Bilal Sidhu. See y'all in the next one.