Alison Gopnik, a conversation with Tyler Cowen (children as scientist, Freud & Piaget, AI)

Tyler Cowen, Alison Gopnik on Childhood Learning, AI as a Cultural Technology, and Rethinking Nature vs. Nurture (Ep. 265), Conversations with Tyler, Dec. 17, 2025.

The introduction:

Alison Gopnik is both a psychologist and philosopher at Berkeley, studying how children construct theories of the world from limited data. Her central insight is that babies learn like scientists, running experiments and updating beliefs based on evidence. But Tyler wonders: are scientists actually good learners? It’s a question that leads them into a wide-ranging conversation about what we’ve been systematically underestimating in young minds, what’s wrong with simple nature-versus-nurture frameworks, and whether AI represents genuine intelligence or just a very sophisticated library.

Tyler and Alison cover how children systematically experiment on the world and what study she’d run with $100 million, why babies are more conscious than adults and what consciousness even means, episodic memory and aphantasia, whether Freud got anything right about childhood and what’s held up best from Piaget, how we should teach young children versus school-age kids, how AI should change K-12 education and Gopnik’s case that it’s a cultural technology rather than intelligence, whether the enterprise of twin studies makes sense and why she sees nature versus nurture as the wrong framework entirely, autism and ADHD as diagnostic categories, whether the success of her siblings belies her skepticism about genetic inheritance, her new project on the economics and philosophy of caregiving, and more.

Kids as Scientists:

We have some good computational models of how scientific theory change works. It turns out that those apply to children as well. The specific thing that I’ve looked at is, what is it that scientists do? Here’s this big, hard problem. All we seem to get from the world are a bunch of photons at the back of our retina and disturbances of air in our ears, and yet, children know about people and things, and scientists know about quarks and quantum phenomena. How do we ever get from the data to the theory?

One subcategory of that is, how do we ever get causal structure which is so important in science? How do we ever figure out what causes what just from a bunch of data that we have?

What’s happened is that philosophers of science and computer scientists have found some systematic ways that you could talk about that. Scientists — I think, mostly, not necessarily consciously, but just as part of what they do — and little kids are looking at data and systematically figuring out what kind of structure out there in the world could have caused this pattern of data. That’s not the only thing, of course, that’s going on in science. There’re lots of other things, too, but it’s at least one central thing going on in science that we’ve started to really understand. [...]

If you asked a three-year-old, “Do you think that this pattern of conditional dependencies is giving you a confounding causal structure?” They would probably not give you a very sensible answer. Even when you ask scientists that, they don’t give you a very sensible answer. But when you look at their actual practice, what you see is that, in fact, kids, for example, are Bayesian, and so are scientists.

Now, the thing is that, in fact, in many respects, kids are better Bayesians than scientists, but a lot of it depends on your prior. If you have, as they say, a very peaked prior, you have a lot of experience, you have a lot of reason to believe that this prior assumption is right, then it’s rational not to change it when you just have a little bit of evidence. You should require a lot of evidence to overturn something that you have a lot of confirmation for.

It’s interesting that the kids, actually, are better at solving problems that involve unusual outcomes than the scientists are. I think what happens in science — we’ve just been doing some work about this — is that there’s also a social factor, where having a big distribution of people who are more likely to go with the prior versus people who are more likely to go with the evidence, which seems to be true in science, that collectively can get you to the right answer. There’s no arbitrary principle you can have about when should you abandon the theory and when should you hold onto it. [...]

One thing you can do, which is like what you’re describing about the money supply, is just make little changes to what you already know. That’s what you mean about moving in the predictable direction. You’re just changing things a little bit. Then seeing, “Okay, if I change it a little bit, is it doing a better job of accounting for the data?” That’s what people think of as a low-temperature search. The other kind of search you can do, the high-temperature search, is just bounce around the space. Try wild, crazy things. Exactly as you were saying, have just a more random walk.

The strategy that you see in computer science, this annealing, is start out with this wild, crazy, out-of-the-box, high-temperature search through the space, and then cool off and just fill in the details. If you think about your four-year-old, who do they sound like? Do they sound like the creature that’s just moving a little bit, or do they sound like they’re noisy and bouncy and random and doing all sorts of weird things? The four-year-old seemed to be a really good idea of this kind of random search. [...]

I think you see both things happening. When you get big paradigm shifts, as Kuhn said, when you get big changes in science, a lot of times it’s because someone found an idea that looked like it was improbable. The nice thing about kids is, because they don’t have to worry about grant proposals, they can be off in the wild space all the time.

[...]

With scientists, we underestimate how much that — we sometimes dismissively call it a fishing expedition — how much that very general experimentation is playing a role in scientific progress. In the grant, you’re supposed to say, “Here’s my three hypotheses, and here are the four experiments I’m going to do to test them.” But I think in practice, a lot of times, scientists are being like the little boy with the avocado and the spoon. They’re saying, “I don’t know, what will happen if I try this? What will happen if I try that?” Then they write the grant to get money to do the things that they’ve already done by doing all these experiments.

FWIW, I've known about simulated annealing for years, a couple of decades at least. For awhile I was one of my go-to metaphors/analogies, though I've not used it recently. In terms that I've been developing in other posts and in some working papers, high-temperature search is ludic (Homo Ludens) while low-temperature filling-in-the details is economic (Home Economicus).

LLMs tend to be used in economic ways. All those benchmarks are based on specific problems in well-specified domains. That's why they aren't particularly creative. My series of blog posts on humans in the loop contains case studies of three of my own ludic explorations.

LLMs as cultural technologies: Four Views

Henry Farrell, Large language models are cultural technologies. What might that mean? Programmable Mutter, Aug. 18, 2025.

It’s been five months since Alison Gopnik, Cosma Shalizi, James Evans and myself wrote to argue that we should not think of Large Language Models (LLMs) as “intelligent, autonomous agents” paving the way to Artificial General Intelligence (AGI), but as cultural and social technologies. In the interim, these models have certainly improved on various metrics. However, even Sam Altman has started soft-pedaling the AGI talk. I repeat. Even Sam Altman.

So what does it mean to argue that LLMs are cultural (and social) technologies? This perspective pushes Singularity thinking to one side, so that changes to human culture and society are at the center. But that, obviously, is still too broad to be particularly useful. We need more specific ways of thinking - and usefully disagreeing - about the kinds of consequences that LLMs may have.

This post is an initial attempt to describe different ways in which people might usefully think about LLMs as cultural technologies. Some obvious provisos. It identifies four different perspectives; I’m sure there are more that I don’t know of, and there will certainly be more in the future. I’m much more closely associated with one of these perspectives than the others, so discount accordingly for bias. Furthermore, I may make mistakes about what other people think, and I surely exaggerate some of the differences between perspectives. Consider this post as less a definitive description of the state of debate than a one man presentation exchange that is supposed to reveal misinterpretations and clear the air so that proper debate can perhaps get going. Finally, I am very deliberately not enquiring into which of these approaches is right. Instead, by laying out their motivating ideas as clearly as I can, I hope to spur a different debate about when each of them is useful when and for which kinds of questions.

Gopnikism

I’m starting with this because for obvious reasons, it’s the one I know best. The original account is this one, by Eunice Yiu, Eliza Kosoy and Alison, which looks to bring together cognitive psychology with evolutionary theory. They suggest that LLMs face sharp limits in their ability to innovate usefully, because they lack direct contact with the real world. Hence, we should treat them not as agentic intelligences, but as “powerful new cultural technologies, analogous to earlier technologies like writing, print, libraries, internet search and even language itself.”

Behind “Gopnikism” lies the mundane observation that LLMs are powerful technologies for manipulating tokenized strings of letters. They swim in the ocean of human-produced text, rather than the world that text draws upon. Much the same is true, pari passu, for LLMs’ cousin-technologies which manipulate images, sound and video. That is why all of them are poorly suited to deal with the “inverse problem” of how to reconstruct “the structure of a novel, changing, external world from the data that we receive from that world.”

Interactionism

Interactionist accounts of LLMs start from a similar (but not identical) take on culture as a store of collective knowledge, but a different understanding of change. Gopnikism builds on ideas about how culture evolves through lossy but relatively faithful processes of transmission. Interactionism instead emphasizes how humans are likely to interpret and interact with the outputs of LLMs, given how they understand the world. Importantly for present purposes, cultural objects are more likely to persist when they somehow click with the various specialized cognitive modules through which human intelligence perceives and interprets its environment, and indeed are likely to be reshaped to bring them more into line with what those modules lead us to expect.

From this perspective, then, the cultural consequences of LLMs will depend on how human beings interpret their outputs, which in turn will be shaped by the ways in which biological brains work. The term “interactionism” stems from this approach’s broader emphasis on human group dynamics but by a neat coincidence, their most immediate contribution to the cultural technology debate, as best as I can see it, rests on micro-level interactions between human beings and LLMs.

Structuralism

I’ve recently written at length about Leif Weatherby’s recent book, Language Machines, which argues that classical structuralist theories of language provide a powerful theory of LLMs. This articulates a third approach to LLMs as cultural technologies. In contrast to Gopnikism, it doesn’t assume that culture’s value stems from its connection to the material world, and pushes back against the notion that we ought build a “ladder of reference” from reality on up. It also rejects interactionists’ emphasis on human cognitive mechanisms:     

A theory of meaning for a language that somehow excludes cognition—or at least, what we have often taken for cognition—is required.

Further:

Cognitive approaches miss that the interesting thing about LLMs is their formal-semiotic properties independent of any “intelligence.”

Instead of the mapping between the world and learning, or between the architecture of LLMs and the architecture of human brains, it emphasizes the mappings between large scale systems. The most important is the mapping between the system of language and the statistical systems that can capture it, but it is interested in other systems too, such as bureaucracy.

Language models capture language as a cultural system, not as intelligence. … The new AI is constituted as and conditioned by language, but not as a grammar or a set of rules. Taking in vast swaths of real language in use, these algorithms rely on language in extenso: culture, as a machine.

The idea, then, is that language is a system, the most important properties of which do not depend on its relationship either to the world that it describes or to the intentions of the humans who employ it.

Role play

Weatherby is frustrated by the dominance of cognitive science in AI discussions. The last perspective on cultural technology that I am going to talk about argues that cognitive science has much more in common with Wittgenstein and Derrida than you might think. Murray Shanahan, Kyle McDonell and Laria Reynolds’ Nature article on the relationship between LLMs and “role play” starts from the profound differences between our assumptions about human intelligence and how LLMs work. Shanahan, in subsequent work, brings this in some quite unexpected directions.

I found this article a thrilling read. Admittedly, it played to my priors. I first came across LLMs in early/mid 2020 thanks to “AI Dungeon,” an early implementation of GPT-2, which used the engine to generate an infinitely iterated role-playing game, starting in a standard fantasy or science fiction setting. AI Dungeon didn’t work very well as a game, because it kept losing track of the underlying story. I couldn’t use it to teach my students about AI as I had hoped, because of its persistent tendency to swivel into porn. But it clearly demonstrated the possibility of something important, strange and new.

There's much more at the link.

Needless to say, I am very sympathetic to this line of thinking. 

Cultural Technology, Old School (in Jersey City)

Large AI models are cultural and social technologies

Henry Farrell, Alison Gopnik, Cosma Shalizi, and James Evans, Large AI models are cultural and social technologies, Science, 13 Mar 2025, Vol 387, Issue 6739 pp. 1153-1156, DOI: 10.1126/science.adt9819

Abstract: Debates about artificial intelligence (AI) tend to revolve around whether large models are intelligent, autonomous agents. Some AI researchers and commentators speculate that we are on the cusp of creating agents with artificial general intelligence (AGI), a prospect anticipated with both elation and anxiety. There have also been extensive conversations about cultural and social consequences of large models, orbiting around two foci: immediate effects of these systems as they are currently used, and hypothetical futures when these systems turn into AGI agents—perhaps even superintelligent AGI agents. But this discourse about large models as intelligent agents is fundamentally misconceived. Combining ideas from social and behavioral sciences with computer science can help us to understand AI systems more accurately. Large models should not be viewed primarily as intelligent agents but as a new kind of cultural and social technology, allowing humans to take advantage of information other humans have accumulated.

Here's an ungated version.

Transmission Versus Truth, Imitation Versus Innovation: Children vs. LLMs

Yiu, E., Kosoy, E., & Gopnik, A. (2023). Transmission Versus Truth, Imitation Versus Innovation: What Children Can Do That Large Language and Language-and-Vision Models Cannot (Yet). Perspectives on Psychological Science, 0(0). https://doi.org/10.1177/17456916231201401

Abstract: Much discussion about large language models and language-and-vision models has focused on whether these models are intelligent agents. We present an alternative perspective. First, we argue that these artificial intelligence (AI) models are cultural technologies that enhance cultural transmission and are efficient and powerful imitation engines. Second, we explore what AI models can tell us about imitation and innovation by testing whether they can be used to discover new tools and novel causal structures and contrasting their responses with those of human children. Our work serves as a first step in determining which particular representations and competences, as well as which kinds of knowledge or skills, can be derived from particular learning techniques and data. In particular, we explore which kinds of cognitive capacities can be enabled by statistical analysis of large-scale linguistic data. Critically, our findings suggest that machines may need more than large-scale language and image data to allow the kinds of innovation that a small child can produce.

Alison Gopnik on children, exploration, play and AI [R&D at the skunkworks]

Ezra Klein interviews psychologist Alison Gopnik, Why Adults Lose the ‘Beginner’s Mind’, NYTimes, April 16, 2021.

Children as explorers:

Klein: You write that children aren’t just defective adults, primitive grown-ups, who are gradually attaining our perfection and complexity. Instead, children and adults are different forms of Homo sapiens. How so?

Gopnik: Well, from an evolutionary biology point of view, one of the things that’s really striking is this relationship between what biologists call life history, how our developmental sequence unfolds, and things like how intelligent we are. And there’s a very, very general relationship between how long a period of childhood an organism has and roughly how smart they are, how big their brains are, how flexible they are. And an idea that I think a lot of us have now is that part of that is because you’ve really got these two different creatures. So you’ve got one creature that’s really designed to explore, to learn, to change. That’s the child form. And then you’ve got this other creature that’s really designed to exploit, as computer scientists say, to go out, find resources, make plans, make things happen, including finding resources for that wild, crazy explorer that you have in your nursery. And the idea is that those two different developmental and evolutionary agendas come with really different kinds of cognition, really different kinds of computation, really different kinds of brains, and I think with very different kinds of experiences of the world. So, the very way that you experience the world, your consciousness, is really different if your agenda is going to be, get the next thing done, figure out how to do it, figure out what the next thing to do after that is, versus extract as much information as I possibly can from the world. And I think adults have the capacity to some extent to go back and forth between those two states. But I think that babies and young children are in that explore state all the time. That’s really what they’re designed to do. They’re like a different kind of creature than the adult. You sort of might think about, well, are there other ways that evolution could have solved this explore, exploit trade-off, this problem about how do you get a creature that can do things, but can also learn things really widely? And Peter Godfrey-Smith’s wonderful book — I’ve just been reading “Metazoa” — talks about the octopus. And the octopus is very puzzling because the octos don’t have a long childhood. And yet, they seem to be really smart, and they have these big brains with lots of neurons. But it also turns out that octos actually have divided brains. So they have one brain in the center in their head, and then they have another brain or maybe eight brains in each one of the tentacles. And if you actually watch what the octos do, the tentacles are out there doing the explorer thing.

States of consciousness:

Klein: And is that the dynamic that leads to this spotlight consciousness, lantern consciousness distinction? And can you talk about that? Because I know I think about it all the time.

Gopnik: So those are two really, really different kinds of consciousness. One kind of consciousness — this is an old metaphor — is to think about attention as being like a spotlight. It comes in. It illuminates the thing that you want to find out about. And you don’t see the things that are on the other side. And I think that in other states of consciousness, especially the state of consciousness you’re in when you’re a child — but I think there are things that adults do that put them in that state as well — you have something that’s much more like a lantern. So you’re actually taking in information from everything that’s going on around you. And the most important thing is, is this going to teach me something? Is this new? Is this interesting? Is this curious, rather than focusing your attention and consciousness on just one thing at a time. [...] think about when you’re completely absorbed in a really interesting movie. You’re kind of gone. Your self is gone. You’re not deciding what to pay attention to in the movie. The movie is just completely captivating. In the state of that focused, goal-directed consciousness, those frontal areas are very involved and very engaged. And there seem to actually be two pathways. One of them is the one that’s sort of here’s the goal-directed pathway, what they sometimes call the task dependent activity. And then the other one is what’s sometimes called the default mode. And that’s the sort of ruminating or thinking about the other things that you have to do, being in your head, as we say, as the other mode. When you look at someone who’s in the scanner, who’s really absorbed in a great movie, neither of those parts are really active. And instead, other parts of the brain are more active. And that brain, the brain of the person who’s absorbed in the movie, looks more like the child’s brain.

Play:

Klein: Do you think for kids that play or imaginative play should be understood as a form of consciousness, a state?

Gopnik: Yeah, that’s a really good question. So there’s really a kind of coherent whole about what childhood is all about. So if you think from this broad evolutionary perspective about these creatures that are designed to explore, I think there’s a whole lot of other things that go with that. So one thing that goes with that is this broad-based consciousness. But another thing that goes with it is the activity of play. And if you think about play, the definition of play is that it’s the thing that you do when you’re not working. Now it’s not a form of experience and consciousness so much, but it’s a form of activity. It’s a form of actually doing things that, nevertheless, have this characteristic of not being immediately directed to a goal. If you look across animals, for example, very characteristically, it’s the young animals that are playing across an incredibly wide range of different kinds of animals. Sometimes if they’re mice, they’re play fighting. And if they’re crows, they’re playing with twigs and figuring out how they can use the twigs. So, what goes on in play is different. But it’s really fascinating that it’s the young animals who are playing. And all of the theories that we have about play are play’s another form of this kind of exploration. So it’s another way of having this explore state of being in the world. [...] 

...children are the R&D wing of our species...

Klein: I was thinking about how a moment ago, you said, play is what you do when you’re not working. And I was thinking, it’s absolutely not what I do when I’m not working. I’m constantly like you, sitting here, being like, don’t work. And that’s not playing. And in fact, I think I’ve lost a lot of my capacity for play. I’ve trained myself to be productive so often that it’s sometimes hard to put it down. And it takes actual, dedicated effort to not do things that feel like work to me. What’s lost in that? Because I think there’s cultural pressure to not play, but I think that your research and some of the others suggest maybe we’ve made a terrible mistake on that by not honoring play more.

Kids on consciousness and mind

Over at Edge David Chalmers is holding forth on consciousness. I believe he's the one responsible for foisting the notion of a "hard problem" of consciousness on him. I think that notion is over-rated, but that's neither here nor there. I'm interested in some remarks that come up during the discussion.

[Rodney] BROOKS: What age do kids start reporting on consciousness? Do you have any idea?

CHALMERS: It depends where you count. Are you talking about consciousness in general, the abstract category? This comes relatively late. What age do kids start talking about pain?

ALISON GOPNIK: If you’re talking about things like differences between mental states and physical states, by the time kids are three they’re saying things like, "If I’m just imagining a hotdog, nobody else can see it and I can turn it into a hamburger. But if it’s a real hotdog then everybody else can see it and I can’t just turn it into something else by thinking it." There's a bunch of work about kids understanding the difference between the mental and the physical. They think that mental things are not things that everybody can see, and that you can alter them in particular kinds of ways, whereas physical things can't, and that’s about age three or four.

There is a whole line of research that John Flavell did, where you ask kids things like, "Ellie is looking at the wall in the corner, are things happening inside of her mind?" It’s not until about eight or nine, until late from a developmental perspective, that they say something’s going on in her mind when she’s sitting there and not acting.

You can show that even if you give the introspective example; for example, if you ring a bell regularly—every minute the bell rings—and then it doesn’t, and you say to the kid, "What were you thinking about just now?" The kids say, "Nothing." You ask them if they were thinking about the bell and they just say no. There’s a lovely passage where a kid says that the way your mind works is there are little moments when something happens in your mind, you think, and then nothing happens in there. Their meta view is that it’s consciousness if you’re perceiving, or acting, or imagining to a prompt. But if you don’t, if it’s not connected, then nothing is happening. So, they have a theory of consciousness, but it looks like it’s different. [...]

GOPNIK: Here’s a proposal, David, that’s relevant to kids not wanting to go to sleep. One of the things that’s very characteristic of kids, including babies from an early age, is that at a point when they clearly have an incredibly strong drive to go to sleep, they don’t want to go to sleep. If you talk to kids, even little kids, it’s very hard not to conclude that the reason they don’t want to go to sleep is because they don’t want to lose consciousness. It’s sort of like, "I’ve only been able to do this for two years, I really don’t want to stop." I don't know whether other creatures share that.

CHALMERS: That's an intuition about the idea of consciousness, that it does something special that gives your life value.

GOPNIK: Nick Humphrey has an interesting proposal along these lines that it’s connected to things like not wanting to die, that that's the reason for the meta-intuition.

CHALMERS: So, he thinks that actually generates the problem of consciousness, because we don't want to die.

FRANK WILCZEK: We know we go to sleep, but we’re not so sure we’re going to wake up.

Bonus, how anesthesiologists think about consciousness:

SETH LLOYD: I had a conversation about consciousness with an anesthesiologist and she pointed out that if you’re an anesthesiologist, consciousness is definitely not one thing because you have to have four different drugs to deal with the different aspects of consciousness that you wish to disable. You have one to just knock people out. It's known that people can still experience things and still experience pain, so then you have another to block the sensation of pain. People could still have memories while they’re knocked out and not feeling pain, so you have to give them another one to knock out the memories that you have. Sometimes they give you an extra special one to make you feel good when you wake up. So, each of these drugs are quite different from each other, with different functions, and they’re disabling different aspects of the things that we call "consciousness."

And then we have Sydney Lamb's daughter. Lamb opens his book, Pathways of the Brain, with with this anecdote (p. 1):

Some years ago I asked one of my daughters, as she sat at the piano, "When you hit that piano key with your finger, how does your mind tell your finger what to do?" She thought for a moment, her face brightening with the intellectual challenge, and said, "Well, my brain writes a little note and sends it down my arm to my hand, then my hand reads the note and knows what to do." Not too bad for a five-year old.

Lamb goes on to suggest that an awful lot of professional thinking about the brain takes place in such terms (p. 2):

This mode of theorizing is seen in ... statements about such things as lexical semantic retrieval, and in descriptions of mental processes like that of naming what is in a picture, to the effect that the visual information is transmitted from the visual area to a language area where it gets transformed into a phonological representation so that a spoken description of the picture may be produced....It is the theory of the five-year-old expressed in only slightly more sophisticated terms. This mode of talking about operations in the brain is obscuring just those operations we are most intent in understanding, the fundamental processes of the mind.