Speaking of Psychology - How children's amazing brains shaped humanity, with Alison Gopnik, PhD

Episode Date: February 17, 2021

As a species, humans have an extra-long childhood. And as any parent or caregiver knows, kids are expensive—they take an extraordinary amount of time, energy and resources to raise. So why do we have such a long childhood? What’s in it for us as a species? According to Alison Gopnik, PhD, of the University of California, Berkeley, the answer is that kids are the "R&D division of humanity," with brains optimized to explore the world and seek out new knowledge and experiences. Gopnik discusses her research and its implications for how we think about the purpose of childhood, how we raise and educate our children, the role of grandparents in teaching the next generation, and even how we might develop artificial intelligence systems inspired by children’s remarkable learning abilities.  Are you enjoying Speaking of Psychology? We’d love to know what you think of the podcast, what you would change about it, and what you’d like to hear more of. Please take our listener survey at www.apa.org/podcastsurvey. Links Alison Gopnik, PhD Music Inspiring Dramatic Beat #07 by tyops via Freesound.org Learn more about your ad choices. Visit megaphone.fm/adchoices

As the crispy chicken sandwich from 7-Eleven, people always call me loud. And I'm like, yeah, I know. I'm crispy. Did you expect me to whisper? If you want quiet, go eat some soup and reflect. Like, I know I'm a handful. I'm bold, I'm juicy. Throw some pickles and barbecue sauce on me, and baby, I'm a whole meal. And with seven rewards, I'm just $4.

Quiet. No. Crispy, saucy, and $4? Very. Only at 711. Valley 36-2326, participating stores only while supplies lastly out for full terms. As a species, we humans have an extra long childhood. Our closest primate relatives are more or less self-sufficient by about age seven,

nearly a decade before human kids are ready to face the world on their own. And as any parent or caregiver knows, children are costly. They take an extraordinary amount of time, energy, and resources to raise. Children are not simply messier and less sophisticated versions of adults either. A baby's brain has unusual capabilities. It's optimized to explore the world and seek new knowledge and experiences. Knowing this has implications for how we think about the purpose of childhood and how we raise and educate our children, even how we might develop artificial intelligence inspired

by children's remarkable learning abilities. So why do we humans have such a long childhood? What's in it for us as a species? What can we learn about our adult cells by looking at kids' brains? And is there anything we adults can do to become more open, curious, and childlike in our thinking. How might that be beneficial? Welcome to Speaking of Psychology, the flagship podcast of the American Psychological Association that examines the links between psychological science and everyday life. I'm Kim Mills. Our guest today is Dr. Allison

Gopnik, a professor of psychology and an affiliate professor of philosophy at the University of California, Berkeley. She studies how children learn and come to understand the world around them. Over her four-decade career, she's published more than 100 journal articles and many books, including her most recent The Gardener and the Carpenter, what the new science of child development tells us about the relationship between parents and children. She's also written widely about topics including play, schooling, the relationship between grandparents and grandchildren, artificial intelligence, and much more that we'll try to touch on today if we can.

Thank you for joining us, Dr. Gopnik. Happy to be here. You have written that kids are the R&D, the research and development division of humanity. That's a really great image. Can you tell us what you mean by that? Well, one of the interesting puzzles, as you mentioned, is why is it that we have such a long period of childhood? Why is it that, you know, not to put too fine a point on it, our babies are useless. In fact, you could argue they're sort of worse than useless because we have to put so much time and energy into taking care of them.

And that's a kind of evolutionary puzzle. And it turns out that if you look across many different species, you see this relationship between how long a period of childhood an animal has and essentially how smart, how good they are at learning as adults. So you see this not just for humans, but you see it among primates. You even see it among birds. So that, for instance, ravens and crows are very, very smart birds. And they have a much longer period, fledgling periods than other. less intelligent birds. So that raises the question,

what is it that that period is doing? Was there some function it's serving that makes it worth the cost of taking care of those helpless babies? And the best explanation we have is that there's a kind of trade-off between different kinds of things that a brain or a mind can do. In artificial intelligence,

they talk about this is the explore-exploit trade-off. And the idea is that you can't really have a system that's both really, really, really, at doing things and really, really good at learning things. Because the things that you need to be effective at acting, having very narrow focus, acting very swiftly, only considering a couple of options at a time, those are all the things you need to,

you know, sort of be an executive, go out and make things happen. And they're very different from the kinds of things that you need to be able to take in as much information as you can from the world, consider as many different possibilities as you can, go out and explore and take risks just for the things, sake of getting information. Those two things are really genuinely in tension with one another. And what I've argued is that childhood is evolution's way of resolving that tension. So childhood gives us this period where we can we can just explore. And the reason why we can do it is

because the one thing that babies are really good at is being as cute as they could possibly be and that that makes everybody else around them want to put in the time and work and investment of taking care of them. And then that period let's let, let, let, let, an animal explorer learn in a kind of free-ranging way. And then, of course, as an adult, you can take the things that you learned in that childhood period and actually put them to use to be able to do all the kind of planning and action and executive function that we do as adults. What do you think it feels like to be a kid and have your mind in this exploratory state

all the time? Is there anything that we experience in adulthood that might come close? Yeah. So it's a really good question about, you know, what is it like to be a base? There's the famous paper by the philosopher Thomas Nagel about what is it like to be a bat and what is it like to be a baby. And again, I think part of the big picture of the point of the work that I've done is it's very easy for people to think about babies as if they're sort of defective grownups. The way to think about development is, okay, there's this absolute marvelous peak of human development, which is, you know, being, I don't know, the 30-year-old philosopher who's writing the article. And then everything else is just sort of building up to that or else falling off from that later on when you say become an old lady the way I am now.

But of course, from an evolutionary perspective, that doesn't make very much sense, right? And you can see that in discussions about consciousness too. So there's been this sort of idea that, you know, the 35-year-old philosopher in his armchair is the peak of consciousness. And when he's sitting there thinking about consciousness, that's the greatest kind of consciousness you could have. So that means that babies who are obviously not doing that sort of thing must be less conscious than adults. But I think there's a good argument we made that babies are actually more conscious than adults. And the reason is if you think about what the function of their, the functional characteristics of their experience, the fact that they're learning a lot, the fact that everything around them is new,

even the fact that when you look at their brains, there's a tremendous amount of plasticity, as neuroscientists say, a lot of change going on. All that fits with a picture that they're actually more open to more of what's going on in the world around them than we typically are when we're adults sort of stuck in our little narrow internal paths. And I think they're, you know, for instance, the kind of classic Buddhist psychology

has pointed out this contrast between the way we think when we're just into our own heads and focused on one specific thing, and the way that in some kinds of meditative states, like open awareness states, for example, we're suddenly open to everything that's going on around us. We're aware of everything around us. I think another example is travel for adults. So when you go to a new place, you are not very functional.

You don't, you often, you know, at least like me, you leave your, you know, you leave your, your charger in five different places. and you're not very stressful to even get from, you know, your hotels at the conference center. But on the other hand, you're you're much more aware of everything that's going on around you. And, you know, there's a couple of days in the new place in Beijing or Paris feel as if they're full of consciousness in a really deep, important way that it's not like your everyday experience. And then the most recent thing that I think might be revealing about this is that there's some recent work that suggests, that people under the effects of certain kinds of psychedelic substances, that there's been a great deal of work recently about kind of renaissance of work

on psychedelics and their potential effects. And the brain effects of those looks like from the new studies that the brain effects of those substances are a lot like putting the brain back into its childhood state, the state of a great deal of plasticity and not very much control, not very much frontal control. And again, that sort of, you know, if you sort of think about, I think everybody who's seen a baby has had that experience of thinking, gee, that baby is just really out there, right? You watch the baby looking at their hands or looking at their mobile and thinking this is the most amazing thing in the world.

Everything's a discovery. Sort of fit. That's right. So that the babies seem to be more tuned into even. little details of the world around them, even though they don't have the kind of focused internal, all right, here's what I'm going to do tomorrow, here's what I'm going to do the day after that, that we do as adults. You've talked about how children's brains are lanterns of consciousness, while adult brains are like spotlights of consciousness. That's what you're talking about here,

right? And what do you mean by that? Exactly. So the metaphor about attention, for example, being a spotlight, that's a very old metaphor in cognitive psychology. And there's lots and lots of evidence to suggest that that's true. So for adults, so there's even, your listeners probably know about the famous guerrilla experiment where there's a phenomenon called inattentional blindness, which means that we as adults literally don't see things

that we're not paying attention to. So if someone says, okay, here's, you know, two people and they're throwing balls back and forth, and I want you to concentrate on how many times did the ball move, and then a gorilla walks right through the middle of the screen, people literally don't see the gorilla, which is sort of amazing. But it gives you a sense of how narrow and bright that spotlight is. And I think if you look at babies and young children, you're not seeing that kind of inattentional blindness. The focus of what the babies are doing is much

wider. And of course, that has some disadvantages, like it's hard for babies to pay attention or to concentrate. But, you know, one of the things I say is when we say that children are bad at paying attention. What we really mean is they're bad at not paying attention. So what they're bad at is filtering out all the other kinds of things that are going on around them, that fascinating little piece of dust on the floor or the sound of that airplane high up in the sky that we grownups don't even notice. They're bad at not paying attention to those things and just paying attention to the thing they're supposed to be paying attention to, like putting on their jacket to get to school. In recent years, you've become interested in the other end of the

the lifespan as well. I know you've written about in addition to our having long childhoods, that humans also have unusually long old age. And you believe that both of these factors are critical to human intelligence and adaptability. Why is that? So there's two things that are that are really interesting about this. One is, as we know, humans have this long, essentially grandmotherhood, this post-manipausal period of about 20 years, which, really makes us different from our closest primate relatives and makes us different from for most other animals. And that's always been true. That was true even when we were foragers that women live for 20 years after menopause. And of course, that's also true for men. So men live for that extra 20 years as well.

It's not quite as vivid an evolutionary paradox with men, but it's true in both cases. And that happened pretty quickly. So then there's this puzzle about why do we have that extra 20 years there? And one piece is that, and you know, that 20 years comes with decreased strength and speed and all the things that we know about as we as we get older. And one piece clearly is that those immature, intelligent children need a lot more parental investment. They need a lot more taking care of than, say, primate babies do. And there's some very nice evidence in anthropology that it's those older people who are actually putting in the time and energy and investment to just help keep those children alive.

But there's an even more interesting story. The one creature that we know of that also has post-metapausal grandmothers, aside from humans, are orcas. Killer whale. It might seem a bit surprising. Why orchids? Well, it turns out orcas are among the few other species that also has cultural traditions where a particular pod of orcas will choose a particular kind of food, you know, krill or or shrimp or other fish.

And it turns out that it's actually the grandmothers who are leading the pod who are pointing to the sources of the food and and seem to be responsible for some of this cultural transmission. Now, one of the things that I think has become increasingly clear about, human is that we depend on these cultural traditions again much more than any other species other species have these traditions but it's you know it's sort of the secret of our success it's one of the things that's most distinctive about us if you think about all the differences between what we're doing now as we sit here on our our Zoom computers and what we were doing not so long ago in the

Pleistocene all of that's due to cumulative culture and if you think about elders like grandmothers one of the things that they can do is pass on information about what people have discovered in the past. So before writing the 150 years between my grandchildren and my grandparents, I can tell them stories about my own grandparents, that's about as far back as you could go. And there's some fascinating anthropological evidence that that's one of the main things that elders are doing. So a lot of this work is in a special issue of the philosophical transactions of the Royal

Society that I edited recently and there's a marvelous piece by Michael Gervyn who was looking at hunting. And what he points out, this will be familiar to everybody in COVID time, is that it's quite hard to actually do something well and teach someone else how to do it well. So when we make pancakes in my house in the wonderful pre-COVID days with my grandchildren, one of the discoveries that I made is when the grandchildren weren't there, it's so much faster to make the pancakes. It's like four times. four times as fast. On Sunday mornings I keep saying to my husband, how come this is going so quickly?

And I think everyone who has children at home is aware of that. And what Gervin found was that the hunters, sort of the prime of life, 35-year-old hunters, were the ones who were most effective. But they weren't teaching as much as the elders were. So the elders weren't as good at actually doing the hunting anymore. They weren't as strong, but they were skilled.

They were the ones who had. have the knowledge. So what happened was that those older hunters were going out with the kids and actually teaching them, showing them what was going on, helping them to figure out what was going on in the world. Now, my own personal, this is just a just a so story, but I think a kind of slightly comforting just so story for us oldsters is you think about that function. Another thing that Gervin found was that the elders were the ones who were passing on the stories and the songs and the big structure of how the world worked. And again, if you think about what you do as a grandmother, that's exactly what you do, right?

Tell stories about when you were a kid and sing songs. And if you think about it from a cognitive perspective, again, I think something that's really important that we haven't emphasized enough in psychology is this idea of tradeoffs, that we can't have a system that does everything perfectly, or even does everything well. There's always tradeoffs between some things and others. So I think there might be a trade-off between being able to remember what happened yesterday, right, which is not all that helpful if you're trying to teach a new generation. They know what happened yesterday. But remembering what happened, you know, 50 years ago, that might actually really be helpful.

So that's slightly comforting when for us, people who can't remember where we put our glasses yesterday, but are really, really good about telling stories about what happened when we were a little girl. Yeah, back in the day. So we humans also have a long adolescence. And research shows that when you look at things like risk-taking behaviors, that our brains don't fully mature until our mid-20s. I'm just wondering, are teen brains more like children's brains or adults' brains? Yeah, this is a really fascinating piece. And in a piece that I did in Piano Acid in 2017, really for the first time in my career, I started looking at adolescents and thinking about adolescents. And what happened was one of those nice empirical things where I wanted to, you know, this argument about the kids being more exploratory isn't just sort of made up out of whole cloth. We've done a whole bunch of studies that show that children are actually better at learning than adults in cases where the hypothesis is sort of unusual or in cases where you have to sort of take a risk to be able to get the information that you need. And so we wanted to see how that unfolded over the lifespan. So we tested kids from people from five until up through adulthood. And what we discovered somewhat surprisingly was especially if it was a social task,

a task about how you understand other people. The adolescents, what you saw is this really interesting curve where the preschoolers were very flexible. And then they became less flexible as they got to school age. And adults were not flexible at all. But the adolescents actually had this new peak of flexibility. And that fits with a lot of the neuroscience that suggests that especially in domains that have to do with the social world, you see this kind of reintroduction of plasticity and risk taking in adolescence. And, you know, if you think about it just even intuitively, any of you who've raised a child, right? You know, they're these crazy, difficult, imaginative preschoolers.

And then they turn into, like, I am convinced that a nine-year-old is like the sanest person on the face. of the universe, right? Like nine-year-old, my, my oldest grandson is nine now. And it's just, you know, he's calm, he's sane, he's rational, he's thoughtful. And then they hit adolescents and suddenly, wait a minute, it's like there's three again. They're doing all these things because you told them not to and they're crazy and they're taking risks. And I think there's actually some evidence to suggest that adolescence is like this secondary period of exploration, especially about things like how the social world works, how the social world should be organized. And a lot of the risk taking that we see both in preschoolers and in adolescence

is connected to this explore, exploit, tradeoff about trying to get more information versus taking risks than having costs. You've talked about how your experiences as a sister and as a mother influenced your research and your career path. Can you just tell our listeners a little bit about your own history and how you became interested in studying children?

Because that wasn't considered a serious field of study at the time when you began, right? Well, you know, it's interesting because I actually began my career as a philosopher, and I'm still an affiliate in philosophy, and I still think that's kind of what my real vocation is.

That's the thing that I'm best at. And I was interested as a philosopher, in some of these big questions about how the world works. And the deepest question, the one that I was and continue to be really the most interested, the one that's really really propelled my entire career. Part of the reason why I'm interested in AI now, for example, is how is it that we could ever know about the world around us? So all that gets to us from the world are a bunch of photons hitting the back of our

eyes and disturbances of air at our ears. And yet we know about a world that's full of objects and people and thoughts and eventually, you know, quarks and distant planets. And it always seemed kind of remarkable to me. That's a deep philosophical question going all the way back to Plato and Aristotle. How does that work? How is that possible? But I was also the oldest of six children, which means that I was sort of a caregiver. And I had my first child when I was 23.

So really, my whole life is very few exceptions. I've been, I've been, and thank God for grandchildren. I've been engaged and involved with children. And it always seemed to me that children were the sensible people to look at if you wanted to ask that big question. Because they're the ones who are actually doing all of that learning. They're the ones who are figuring out how the world works. And that, you know, wasn't unique. Pige, the great founder of Covenous Development thought that.

In fact, Socrates thought that looking at children was a way that you could answer those questions. But it was still, you know, for obvious reasons, children are, were and are, you know, stuff that kids, moms take care of. It's women's work. It's domestic. It doesn't quite have the kind of status and seriousness often that other, that other, other kinds of research have. So my sort of, I mean, literally from the time I was just starting out my career very intentionally, I set out. had a sort of choice point of should I just go and become a philosopher and be a woman who's doing the kinds of things that analytic philosophers talk about or should I embrace this very female enterprise of looking at children and just make people realize that looking at children was just as deep, just as interesting, just as significant as looking at brains or looking at computers or or doing reaction. or any other kinds of things that adults do.

I'm not sure I've completely succeeded at that, but that's sort of the mission. And I hope that my students are carrying on that mission. Has the fact that more women are entering research psychology contributed to the changes that you've seen in your field? Yeah, I think there's no doubt about that. So both the fact that more women are involved and the fact that more men are,

the fact that more men are involved with children, right? So one of the puzzles, for instance, you know, the really big revolution has been our understanding of babies and young children. So even when I started out, which is a while ago, but not that long ago, I remember people telling me when I was at Oxford that babies were decorticate, that they were like, you know, kind of like vegetables, that they didn't have a cortex. They couldn't really think. And even Pige, you know, the great founder of cognitive development thought that babies were, babies and young children up till about five were irrational and egocentric and solipsistic. And we've just totally turned that on its head. That views just 180 degrees different. So now we think that even the youngest babies both know more and learn more than we ever would have

thought. But an interesting kind of puzzle is like, why were they so wrong for so long, right? I mean, what was it? How could how could all of these people, the smartest people, the philosophers, the psychologists, like psychiatrists, how could they've gotten it so wrong? Well, if they weren't the caregivers, they weren't having the same interaction every day with babies, right? I mean, would that be part of it? Exactly. And I think part of the reason was that the people who were actually spending time with babies did think, wait a minute, I think he recognized me. I think he understands this.

They had a rich sense of it. But of course, they weren't the people who were writing the books. And it's really striking to me now. I have this wonderful new four-month-old grandson who very sadly is often. in Montreal, but my granddaughter and grandson literally every morning send me a video of what the baby's doing. And I've realized, of course, that technology means that I get to see how he's imitating or he's looking or he's engaged. But of course, those are just a few moments during the day. A lot of the time during the day, a baby will be asleep or a baby will be fussy or a baby won't be doing

those things that really demonstrate how smart they are. And it's kind of, if you didn't actually, you know, have the video phone sitting right there so that you could pick it up and do a video at the very moment when the baby's doing it, it would just seem like it was lost. And in fact, you know, again, if I think about it when we were starting out, you know, you wanted to record a baby. You had to take a big, heavy, you know, big heavy reel-to-reel recorder and a box full of equipment and hope that you could get something. And I'm wondering if, you know, the new technology means that we'll start getting more data and appreciating more just exactly what babies can do. Let's talk for a minute about your book, The Gardener and the Carpenter. You talk about

current parenting culture doesn't really mesh with what science is telling us about the purpose of childhood. Can you explain the difference between a gardener and a carpenter when it comes to parenting and what insights parents and other caregivers should be getting from this work of yours? Yeah. So I think the model, it's interesting that the very word parenting really only started showing up in the 1970s in the United States. Of course, we've had mothers and fathers and parents forever. But the idea that parenting is this occupation, this thing that you do, is relatively recent. And I think the other thing that's relatively recent is a generation of people raising children who hadn't actually really, you know, being the way I was, the oldest of six

children, hadn't had much experience of raising children before, but had lots of experience of going to school and working. So they tended to think about raising children as being like going to school and working. It's something that you have a bunch of techniques for. If you just get the right technique and you focus enough, you'll be able to bring about the right outcome. It's not quite clear what the right outcome is, but, you know, something like you'll have the child who goes to Harvard, right? And that picture is very, very much not like the picture that comes from thinking about evolution and development. It's almost the opposite of this explorer exploit tradeoff picture that I've been talking about. That picture is a picture where the whole point is to provide children with a rich, safe environment

in which the children can come out in a wide variety of different ways in which things are very variable and unpredictable. A metaphor I've used to try and describe that is if you think about a carpenter who actually goes out and has a picture in their head about here's the chair that I want to build and here's the wood and here's what I'm going to do. And if I'm very precise and careful, I can make exactly the kind of chair that I want versus a gardener where you have no idea what's going to happen in your garden. Or at least from most messy cottage gardeners like me. But also more deeply, one of the things that we've discovered from ecology is that for a garden or any other ecological system to be successful, the key is for it to be variable, for it to be able to do different things in different circumstances, for it to be resilient. So that, you know, when we have a drought the way we have in California, the succulents are going to do really well. and the British cottage garden flowers are not going to do it as well, but it'll reverse when the range come. So having a lot of variability, a lot of difference is one of the things that make systems

resilient. And the point about childhood is to bring about that kind of unexpected variability in the light of change. So it's just the opposite of what the kind of carpenter parenting picture would suggest. Well, I want to talk for a minute about some of the newer technologies and their role in parenting, and it's not at all uncommon to see parents in malls or restaurants handing toddlers or even babies, mobile phones or iPads to distract them or keep them busy. Is that kind of parenting gardening or carpentering? Can you tell? Well, I think it's important to say that one of the functions of this kind of garden has always been that each generation is dealing with a different universe, a different environment. It's different from the one before. That's one of the things,

go back to the grandparents, right? That's one of the things about our ability to have culture is that not only do we pass on culture from before, but we also innovate. So we also get new cultural discovery. And part of what childhood has always been about, adolescents particularly, I think, but childhood as well is, all right, I'm in a slightly different environment than I was, than my parents are. How do I cope with that environment? And of course, that's very vivid if you're thinking about, if you're thinking about technology now. So I think, again, COVID has made it really striking. I think all of us have had this experience of the children around us being much better at dealing with Zoom than we are. It's almost a cliche about how quickly

children are adapting to this new world compared to adults. But of course, what culture is all about, and that's where the grandmothers come in, is that we always have this, and this again is this point about tradeoffs. We have the way we can do that is because we are passing on the values that we think are important to us, the values that we've discovered, the things that we've found out about the world, we can pass those on to children. And then they're equipped with those values to deal with the new world, which is different from the world that we grew up in. And I think technology is a good example of that. So, you know, I can feel myself, look, standing in front of a screen all day is not a great,

happy way to be. And, you know, having the experience of going off into a corner and reading a book or going out into the garden and playing, those are all really good experiences. And I think I can pass that on to my grandchildren at the same time that my grandchildren can figure out, well, how do you, the puzzle that we all have, How do you integrate that with the reality of the new technology? But no matter what, I think they'll never learn penmanship again, right? I mean, you can't impart that particular knowledge.

Well, I have to say as a child who was, if I do say so myself, was always extremely good at school, did very well in everything growing up except day of penmanship. I am as a slightly, slightly disgruffing person who can't draw and can't write, I'm willing to give that up. So we've touched a little bit on artificial intelligence. I know that's an area where you become pretty deeply involved in the research. And just wondering, what can AI researchers learn from developmental psychologists and how would AI based on children's learning be different from the way that many AI systems

are being built right now? Yeah. So, you know, we were talking before about the effect of having women. in the profession and how it changes the kinds of things that we think about. And I think we've had a really nice dramatic example of that recently with AI, where AI's been very heavily dominated by men and the kind of models of AI learning have been obviously dominated by the people who are doing it. But people in AI, you know, for good diversity reasons as well as for intellectual reasons, have started to pay more and more attention to children. And, and

And of course, a lot of the people who study children are women. So they've been listening to what women have to say about how children learn. And again, a wonderful thing about AI is these are engineers. And engineers are great about just being willing to take everything, anything into account that will make their machines work better, no matter what other characteristics are. And if you think about AI, I think there's been real interest in children in particular, in quite recently in the past few years because the big advances in AI have been advances in machine learning in learning. So the big discoveries have been that we do better if we actually try to have machines that can learn from the world rather than trying to program in what they should know in the first place.

But the kinds of learning that these machines do are pretty limited. So they're, you know, something like here's a giant data set of images and I'm going to tell you label these are cats and these are dogs. And now your job is to say which ones are cats and which ones are dogs. And it's amazing that or you know, here's a game of go and now your job is play a million games ago and then figure out which times what you have to do to win versus lose. And it's amazing that the systems can do as much as they can that way. But the result is that they tend to be quite narrow. So they're good at doing the things that you've trained them to do, but they're not very good at generalizing to new examples.

So you give them a slightly unusual picture of a dog, let's say a video of a dog instead of a picture, and they completely bomb out. You change the rules of the game a little bit, and they can't use the things that they learned before. And that makes them very different from children who can have just one or two examples of something that's new and make a really wonderful, unexpected generalization. An example I give a lot again from my grandson is when he was about

for his grandfather said, gee, I wish I could be a kid again. And Augie said, why didn't you try don't eat any broccoli and don't eat any green beans and don't eat any healthy food? And then maybe you could be a kid again. Now, that's a wonderful, you know, that's my brilliant grandson, but any four-year-olds will give you examples like this, where it's not like anyone had told him that, but of course you can see the rationale behind it, right? If eating vegetables turns you from a kid into a big, strong grown-up,

then you should be able to reverse the process. And that's an example of something where without any explicit supervision, without anybody telling you what the answer is, you think of something that's a really good generalization, a really good creative new idea. And that's one of the things that three and four-year-olds are fantastic at doing. That's part of why it's so much fun to just listen to them talk. So that's one thing that the children can do.

So how is it that they do it? And we think that there's, so I've been part of a big project that DARPA, the defense research group that actually helped invent the internet and modern computers, DARPA is funding to try and see if we could design an artificial system that was as smart as an 18-month-old, say. And the things that we think are really crucial that makes the children different from current chaos is that they build models of the world.

They're not just looking at statistics. They're actually thinking of things like, okay, how does eating turn you into a grown-up? Importantly, they explore. So instead of being sort of stuck in their main frames, you know, we children are out there doing experiments all the time. When they do it, we call it getting into everything. But they're getting into everything. They're going out and actually getting information. And there's a lot of exciting work now in

AI trying to make curious AI, that's part of what we've been doing at Berkeley. Could we make AIs that are curious the way that children are? And it looks as if that makes them much smarter than when they're not curious. And then a third thing is that this cultural piece that we learn from other people. So children are imitating, you know, literally from the time they're born. And they imitate in very smart, sophisticated ways. It isn't just kind of mindless copying. They're thinking a lot about what is it that the other person around me is trying to teach me. The work that I did in the first part of my career back in the 80s was what came to be called theory of mind.

How do we understand what's going on in the minds of other people? Again, when we started in the 80s, there wasn't even a word for this. Nobody quite thought of it as being something that would be interesting to find out about. Now it's kind of an industry. And one of the things that, again, people in AI have realized is if you want an AI that works like a person, it needs to understand people. It needs to have a theory of mind. And it needs to use that to try to learn from other people. So can you have a robot that could see someone do something?

And it's sort of fascinating. It's been a wonderful learning experience for me. You know, if you talk to people in robotics, for instance, and you think, okay, well, this should be easy, right? You'll just show the robot that someone's reaching for something and it'll imitate. But it turns out what the robot will do is imitate exactly the absolute trajectory of what the person does. And then if you just move the object over an inch, it'll fail because it doesn't know what it's supposed to what it's supposed to do. Whereas the kids will figure out what is it that you want me to do? Is that the right kind of thing to do? What would be a better way of doing it? They're they're doing this in a much more sophisticated way.

So that's another big piece, especially in robotics that that people are trying to do based on based on what the kids are doing. So we both are finding out more about how kids brains and minds work by trying to say what program are they using that would enable them to learn so much. That's been the big question that I've been asking all along. But that also means that we might be able to actually design some systems that are better at learning.

And then the last thing, which is the most recent thing that I've been working on that I've been most involved in over this last year, year is part of the reason also why children can learn so much is because they have these caring adults around them who are taking care of them and also giving them the kind of information they need to be able to learn. And I'm getting more and more interested again, maybe this is partly being a grandmother, what is involved in being that kind of a caring adult? And it turns out that that's actually very important, oddly enough, in AI, because one of the big problems in AI, it's called the

alignment problem is how could we get AIs that had the same values that we do? So, you know, the big fear about the AIs are going to kill us, which is probably not going to happen. But there's still a question about how do you tell an AI this is what you're trying to do. And it turns out, you know, with humans, it's very hard to actually articulate what is important. What's the value that what's the value that we're trying for? You know, the famous example is like you tell an AI to make paper clips. Does it turn everything in the world into a paper. It's people in AI call it the paperclip apocalypse, right?

Does everything turn into a paperclip? Because you haven't told it, no, no, no, just these things turn into paper clips. And actually, if you think about parenting or being a parent, we face that problem all the time, right? So we want our children to be able to develop new values, develop new goals that aren't the same as ours, but that are good, that are important,

that are not beneficent, right? beneficent rather than negative. So this kind of cultural trade-off of tradition, grandma telling you the values that she had 100 years ago, and then you having to take them and adapt them to a world of Zoom screens, that's actually a problem that we have with AI as well. If we're going to actually have genuinely intelligent AI,

we're going to have to figure out some way that we can pass on, here's what our values and goals are, here's what's important, here's what's good. And also enable the AI to be able to generalize, to discover and think of goals in a new way. And I think that's one of the main problems in AI ethics is how can we, if we want systems that are, we want them to be autonomous,

we don't want to be sitting there and telling them every single thing that they can do, but we want them to be good and we want them to do the things that we think are important. And actually, that's exactly the problem that every generation of caregivers face. And the more I've thought about it, again, I think COVID has made it very vivid that caregiving is something that is absolutely crucial and central to human cognition and progress and human life.

And yet again, as with children in general, it's very neglected. No one's thought about it very much. The moral philosophers don't think about it. It doesn't fit the kind of models that we typically have in our politics. And yet, you know, that individual elder. care worker who cares about that person who's sick and is willing to, you know, risk their own health and take care of them, even though they're paid nothing and they're treated incredibly, incredibly badly. I think there's something so deep about humankind there that we we haven't

thought about very much. We haven't understood in psychological or cognitive or cognitive terms. Or again, you know, watching this grandchild, this new grandchild and, you know, within three weeks. He was actually early. He was premature. And yet, here's this tiny little creature and here's these grownups who just care more about him than anything else

in the world. That's an amazing human capacity that we haven't thought about very much and we don't understand very well. So that's kind of the thing that's kind of on the cutting edge of what I've been doing with my time over the last year. Well, maybe, and this is kind of a last question,

but I think that you perhaps have answered in many ways in just this last topic that we were discussing. But do you think that there's still anything that could surprise you about children's minds after all these years of studying them? Yeah, well, of course, in detail, this is really true. Every time we do an experiment in the lab, we get surprised. So it's a, you know, I'm used to the fact that we'll do a pilot and I'll say, yeah, I'm positive the four-year-olds are going to be able to do this.

I've got, you know, 40 years of experience, I know, and they do something completely. They do something completely unlike, completely unlike what I would have thought. Or vice versa. My student says, how about we try this? We had an experiment recently trying to see if children could understand this, some very abstract ideas from AI about variable choice where I'd say, no, that, come on, that's something that they couldn't possibly understand. And they're, you know, like the two-year-olds are doing it in minutes. So I think there's no question that they're going to continue to do things that are really surprising that I wouldn't have anticipated, but I wouldn't have anticipated in the first place, which of course is part of why it's so much more fun to study them than it is to study any other creatures. Well, this has been really interesting and a lot of fun, Dr. Gopnik. Really, thank you for joining me today.

Well, it's always great to talk to you and to talk to all the APA listeners. Thank you. You can find previous episodes of Speaking of Psychology on our website at speakingofpsychology.org or wherever you get your podcasts. If you have comments or ideas for future podcasts, email us at speaking of psychology at APA.org. Speaking of psychology is produced by Lee Weinerman. Our sound editor is Chris Kondy. Thank you for listening. For the American Psychological Association, I'm Kim Mills.