Latest / Art of the Question / Is AI Killing Creativity? - Prof. Adam Green - #67

Transcript

Is AI Killing Creativity? - Prof. Adam Green - #67 · Art of the Question

• Julius Moorman: Adam, thank you so much for coming on the show. You are a professor and neuroscientist at Georgetown University, and you've done a lot of work about creativity. And that's also the topic of today's discussion. So we're going to learn about what creativity is, what happens in the brain when we engage in creative thinking. Perhaps how we can leverage that knowledge we have about creativity to perhaps instill creativity in ourselves. And I would also love to talk to you about creativity in the relationship to AI because you've also been vocal about that and have also done some research into yeah, Chat GPT and how it's being used in in paper. So that That's something I thought was was very interesting as well. But before we get into all of it, I think it makes sense to kind of start with your definition of creativity. Well, first of all, thanks, Julius. I'm looking forward to the conversation. and it's it's a big question. That's not not not easing in with with the small stuff. So I have thought a lot about this. and you know, I'm one of the few sort of nerds who are totally focused on on creativity and all of our you that that gets to be my job. I mean we we were just talking about the wonderful things about about being an academic. And, you know, I'm thankful so much of the time that that someone's willing to let me spend my time thinking about questions like like this one that you're asking. And we have as a as a community of scholars, the create you know, sort of creativity research community, put forward a few different definitions. I this is preamble, but I I'll get to the answer to your question. I I've through different endeavors in the last 10 to 15 years been in you know, sort of Position where I'm editing a journal on creativity, and also founding you know, sort of helping to lead a society, the Society for the Neuroscience of Creativity. I say this because it has put me in a place to sort of have a view of the way that different or the different ways that that creativity scholars think about creativity, both in the context of what they publish in a journal. And in the discourse that happens within a society, an academic society. And so one of the nice things about that has been the opportunity to sort of absorb those those different perspectives and different approaches to defining creativity. And so as I come around to the answer, what I'll say is that. That ability to observe and absorb in those roles led me to an understanding that I I think is certainly not dramatic or sort of dramatically distinct from ways that that people have approached creativity in the past, but to really take to heart that when we say creativity, it turns out We're saying different things depending on the usage and the context. So, in particular, it seems really important to me to separate when we talk about creativity, whether we're talking about a product, like an invention, could be creative. Hey, that new that new gadget that that someone created, that's that that that's really creative, right? Okay. That's a that's a product. And we'll come back to to why I start with that one. but we could also say, hey, Julius, that guy, he's he's very creative. So now we're talking about a person. Or when we say creativity. To me, and in many instances, I think the most precise referent is a process. That is to say, creativity is thinking creatively, right? Or expressing creative thoughts. And so that that process level is yet another way to To use the word creativity. And it and it turns out those differences in the ways we use the terms are not trivial. And they really end up corresponding to fundamentally different meanings. So what along with some great colleagues, James Kaufman, Roger Beatty, Yoed Kenneth, who are all people in the in the creativity world sort of thought a lot about this and and began parsing creativity along these lines in ways that we feel can be empirically supported, not just theoretically sort of supposed. And so the the first and most important to us usage of creativity or meaning of creativity to really tackle Is this process level? And the process that's important to us because we're people who study cognition or people who study the brain. If you're looking at a brain, you're never looking at the product of creativity. You're only looking at the process. And so that became challenging because the definition that the field has used for a long time, what is sometimes referred to as the standard definition of creativity, Involves two criteria, novel and useful. Something that's creative is novel and useful. But that definition really refers to a product, a new gadget that somebody invented. That gadget can be novel and it can be useful. Well, what does that mean when we want to understand creativity, the process, right? Creativity thinking creatively, engaging in creativity. Well Essentially the answer has just been, well, that process is whatever came before that product that was novel and useful. And if the product was novel and useful, then we'll call the process creativity. And and that's always been kind of profoundly unsatisfying. It's sort of like telling a developmental researcher that, you know, human development is whatever comes before being an adult, right? So whatever that was, that's development. and we don't actually need to understand it. and so. Given the work that I've, you know, I've had some role in doing, but had much more role in just observing and absorbing in the neuroscience of creativity, which has now become a much more mature field, we felt that we had the grounds to really look at what the core elements of a creative process are. And With that neuroscience evidence in mind and a lot of theory and a lot of cognitive science evidence in mind as well, we put together what we call the process definition of creativity. and so we put that into a paper a couple of years back. and it's called the process definition of creativity. and essentially it boils that process down to Internally directed attention constrained by a generative goal. And I can break that down a little bit too. So it's not to say that you shouldn't pay attention to the world outside of you when we say internally directed attention. You can be inspired by the people around you, by the beauty of a sunset. These are all things that can be part of a creative process. But they are not necessarily part of. A creative of a creative process or of every creative process. And in every creative process, if in fact you are the creator, it has to come from in you. And so what that means is you might have seen the sunset or talked to the brilliant person. But when you create something from that experience, that has to be built from your internal representation of that sunset. Or of the wisdom that that that really brilliant person gave you. You've you have to first represent it internally, and then you have to act on it in some way with your attention in order to create something new. Otherwise, you're a camera, right? A camera can absorb a beautiful sunset or or record a beautiful sunset, or that sunset can pass through it. but it is not a creator. and so the difference between being a receiver and being a creator. Is that internally directed attention. And then with the generative goal, the idea is you're there is some goal where something will be created. Now, whether that goal is actually achieved is not essential to whether the thinking was creativity. And I think that's a another really important distinction between the process, defining what the process is. You can imagine I mean, in the case of development, not every child becomes an adult, but that doesn't mean that that is not a case of development, right? And so not every creative process ends in a product, but it is still creativity. So so we've then used a lot of the the sort of neuroscience evidence to talk about how that looks in the brain and and what we could look for if we're measuring things at a process level. But that's that's my long way around to to getting at one definition of one meaning of creativity. And if I could riff on this for just another moment. James Kaufman, who I mentioned before, he and I have been leading now for a couple of years a project that we call the Creativity Ontology Project. And this is a collaboration of 200 of the top creativity researchers around the world, focused on parsing the meaning of creativity as a product, as a process. And as the attribute of a person, along with other creativity-related constructs like imagination or innovation or problem solving and things that get sometimes get used interchangeably or conflated with creativity, but actually shouldn't. if we want any of these things to mean anything, then they can't all just mean the same thing. and so we've been working on that, and that is that that's been a fascinating process. And you know, collaborating with 200 people on anything is is complicated in all the ways you might imagine, but also illuminating in all the ways you might imagine. and that piece is now commissioned at Nature Review Psychology, and so hopefully it will will be out in the next next few months. and so you you can you can look to that to to sort of understand the field's view of of these different meanings of creativity. I'll definitely be on the lookout for that. And and yeah, that sounds that sounds like an incredibly difficult project, but yeah, also very, very exciting. What I found interesting is you mentioned you've used a few analogies there. Development, you've used a camera, and analogies are vital to your earlier work in actually trying to measure creativity, right? Because Creativity is kind of a broad term. There's so many different definitions that one can use. We can use it to describe a product. And I mean, just the work collaborating with 200 people who each have their own take on this underscores the difficulty of defining creativity. But I think you found a very creative way of making creativity a bit more tangible. Can you maybe explain your study about an analogies and and why it was so useful in making Diz broad term a bit more concrete? Yeah. that was that is work that I've been interested in since you know, since I started grad school, since and I won't say how how long ago that was, but it's it's never ceased to be exciting to me that that thread. And I I one of the things as you s as you're sort of noting that came out of that line of work was an opportunity to take this fuzzy thing that is often kind of criticized as being impossible to pin down or or really difficult to study in rigorous empirical ways, and make it more quantifiable. and so what we did is to use the early predecessors. Of these language models that are now so prevalent, use the semantic embedding spaces that are built into those models. And what that means is these spaces where all the meanings in a language swim around together. And they're placed in relative distance to each other. Based on how frequently they're used together. So in this embedding space, you might put baseball and bat right next to each other, but baseball and cabbage would be far apart, right? And when and that's because you rarely see a sentence where baseball and cabbage are used together. the models we have now are more sophisticated versions of that, but are still in that same general tradition. And I should say we we're now using the the newer models when when we do this kind of work. But the basic idea was: well, okay, if these meanings are far apart, but we connect them, that's the kind of connection that we think of as creative. and that's the kind of connection that tends to unlock some hidden and and often very valuable similarity. that we might be able to use to understand something, to build something, to solve something. I mean, there's endless examples, but the, you know, the sort of analogy between the structure of the atom and the structure of the solar system, you know, these are the sorts of things that that have led to scientific insights. you know, Kepler was was a big proponent, but so is Einstein, so is Steve Jobs. I mean All through the sort of history of discovery, analogy has played an important role for this reason. It it shows us some valuable but somewhat hidden connection. And so without making a claim that that's the entirety of creativity, I think it's pretty we're on pretty secure footing to say that is a creative kind of thinking. And That because we had these language models and these embedding spaces, and we could see and we could put a number on how far apart these meanings are in a language, that gave us a way to quantify the creativity of the connections that were being made. Because if I'm connecting two things that that are far apart in an embedding space through my analogy, well, that analogy is probably more creative. So just to give you an example, if you had something like whiteout, well, let's for the n the newer generation probably doesn't even know what what whiteout was, but let's say it pencil and eraser, right? to memory and amnesia, that's that's a pretty far analogy, right? That the the two things that you're connecting are are far apart. and so th they seem different on the surface, but there's a meaningful connection. So being able to make that analogy is one form of valuable creativity, and it's a form that we identified a way to quantify. And because we could quantify it, we could then ask questions about as that quantity changes, what's changing in the brain? what is quantitatively different in terms of brain activity? and so that that opened up some really really fascinating doors and and and exciting ways to explore this question. And then from that, from that sort of early basis of semantic distance, the field of creativity has now really leaned toward semantic distance. in more sophisticated ways, including and this, you know, our our lab has come right along with it in using more sophisticated models to do this, but essentially to to look at how far apart the meanings, say, in a story are. How many different diverse meanings are you able to piece together in your story, right? if I give you a brick. and I s or an example, like a war if I tell you think of ways to use a brick. I could hand you a brick also, but usually we don't get as tactile as that. if I say, okay, what would you do with a brick and and come up with you know creative ways to to use a brick and and you say, well, it could be used to build a house, or it could be used I could grind it up into fairy dust and it could be part of my Halloween costume, right? Though so we can look at the semantic distance between brick. And whatever it was that you said in order to quantify, was that something, well, brick and house that gets used together a lot? That's not very far. That semantic distance is small. Fairy dust, that seems to be pretty far from brick in common usage. And so that's probably a more creative use of the brick. So that that idea of semantic distance that that we developed for analogical. reasoning has now really become kind of a basis for quantitative but and also automated scoring of creativity across a lot of domains. That's super interesting. And and this also a and please correct me if I'm wrong, but also allowed you That as this process was happening, that you could then study what was actually happening inside the brain and see which parts of the brain were activated during the creative process, which then also allowed you to see what would happen if you were to stimulate this part of the brain and to see whether this would actually enhance one's creativity, correct? That's right. Yeah. And so specifically on analogy, When we had this quantitative semantic distance, okay, how far apart are the meanings? Put a number on that. Okay, what changes in the brain when that number is bigger or smaller, when the analogies are farther apart or closer together? Well, one area that consistently seemed interested in that difference between less creative connections and more creative connections. Is an area right here at the front of the brain called frontopolar cortex. and so we, you know, we did a few studies. We said, hey, you know, this area seems to continue to be kind of interested when when we're increasing the creativity of of the analogies that people are either evaluating or generating. and so we we followed that up to say, well, This might not just be a neural signal that I'm thinking creatively. It might be an opportunity, might be showing us an opportunity for how to help brains think more creatively. So just as you're saying, once we had sort of good evidence that this frontopolar activity was associated with increased creative thinking and and connecting meanings across greater semantic distance, we thought, well, there are some tools that we can use to help this region of the brain become more active when people are thinking creatively. And so we used a a a form of electrical stimulation, transcranial direct current stimulation. And we essentially made it easier for those brain cells in this part of the brain To fire when people were thinking creatively. And we found that when they had that help in ramping up activity in those brain cells, they they got more creative. They they were more successful in generating and evaluating these creative analogies. and of course, when you do that work. You you have this crazy helmet that you're wearing that, you know, zaps you. And so i one thing that you have to be careful about is, well, I've got this helmet on my head. I'm I'm a creative genius. Once I put this on, this is the genius helmet. So you have to do these studies in a way where you put the helmet on and you tell everybody that you're zapping them, but you actually only zap half the people. and so everybody thinks they're being zapped. Everybody thinks they're a genius, but then you have to look at, well, what's different when we really were zapping the people? and and and that's where we saw the differences. So Yeah, I think that's really exciting. And I think, you know, Julius, I think it it speaks to something broader in creativity research, but also in cognitive neuroscience, which is which is my field of training, which which is that we're at a place now where we've learned enough about the brain that we can move in many instances from observation to intervention. it doesn't mean that we have a complete understanding. Of what these brain systems are doing. So in the case of frontopolar cortex, we know from a range of studies that frontopolar cortex is important for integrating information. And so that's true if you're making a decision, and it's true if you're representing something complex, but it's also true if you're, it seems to be true from our work, if you're making a creative connection. The ways that it's doing that, we have some indication that it's you know highly connected to areas of the brain where we store a lot of our memories, a lot of our meanings. but the descriptions are at a pretty broad level. And I make that point to say that it seems like we don't necessarily need to have a complete understanding of a brain in order to help it. or in order to modulate it in deliberate ways. And that is actually very consistent with the way that neuroscience has worked in more clinical contexts. So for example, a lot of the the pharmacology, a lot of drugs, so antidepressants are are a great example. they show a lot of efficacy and we s there's still Quite a lot we really don't understand about their mechanism of action. But we we see two things. We see that they are safe and we see that they work. And we have some broad mechanistic understanding for why that that could and and should be the case. but it allows us to sort of take an off-ramp from trying to understand everything about the brain. And I think the brain is ultimately probably too complex to understand itself. The brain trying to understand the brain ends up being a bit of a dead end. But if we can understand the brain at those levels that are actionable and then see what's safe and what's effective at that actionable level, I think there's a lot of exciting things that are happening in cognitive and neuroscience along those lines. Yeah. And for people that do not have an electrical helmet that turns them into a creative genius lying around the house, can we use this knowledge about That this when this part of the brain is stimulated through electrical signals, can we actually mimic that, or can we do certain actions, maybe change our environment, or go for a walk, for example, or do anything that helps to stimulate something similar and make us more creative? Yes. And it's a it's a great question. so the question has different answers depending on the context, and I know we're gonna get to interacting with AI, but I think there are some pieces of advice that I I think apply In in almost every case. So one very simple piece of advice is to pay attention to what you're doing when you feel creative. And try to do more of that. And the more that you put your your brain into a creative state of thinking, the more familiar it will feel and the better you'll get at at creativity. It's a practiced capability. Like so many other things are. And the most important thing is that you get is that you're doing the thing that puts you into that state. And for some people, that might be listening to music. For some people, that might be sitting in a quiet room. For some people, that might be a crowded cafe. For others, maybe not. So that's one broad piece of advice. But I think when we when we really sort of try to leverage what we know about the neuroscience, and we come back to this idea of internally directed attention as part of a creative process. I think what we what we find is most effective is to be able to think in that internally directed way, but not to force yourself to only be looking inward. And so what I mean by that is essentially a context in which you can balance. Outside attention and inside attention, externally directed and internally directed. So for me and for many people, that balance, one great way to get that balance, and and and this comes, I think, with with a lot of people's just personal experiences of when creative thoughts have have come to them or or they've been able to generate creative thoughts. You mentioned the idea of going for a walk. A walk is a terrific way to do that, or a run. But one thing, one piece of advice that I I tend to give is if you go for a walk or you go for a run, do it in a familiar place, a place that you know pretty well. The reason for that is if you're lost in an in an unfamiliar city, or if you're in the woods in a place that you haven't been before, you need a lot of external attention. You're you gotta figure out where the heck you are and you know how to not. You know, fall off a cliff or get, you know, fall down a manhole, you know, like all of these things that that you have to be more vigilant about with your external attention. If you're going for a walk or going for a run in a familiar place, you know all the turns, you know all the trees, you know all the buildings, wherever whatever it is that your context is. So you can look at those things, and they may be quite interesting to look at, but you don't have to look at those things. you'll do just fine when your attention is focused inward. And so having both engaging stimuli, but also the freedom to disengage from those stimuli because you know where you are and you're in a familiar setting. that can allow the sort of balance between internal and external attention that seems to really be a booster or sort of a a a nice likely precondition for for thinking creatively and and and and and thinking well. so I I would say what whatever it is that allows you to have that balance is probably a a good a a good thing to do. but going for a walk or going for a run in a familiar place is is sort of a tangible kind of tip. Yeah, and driving is up there too. And again, mean not not when I drive through the city during rush hour and I'm stressed and I have to look and focus on the road. It is always when you're on the highway, like you say, on a familiar piece of road, when the thoughts usually come pouring in or even standing under the the shower, very famous example as well. It seems to be something with The external stimuli of the water hitting your skin, the sound of the water trickling down, but not really having much else to do, but be in your mind. I think those are scenarios where, at least in my case, a lot of my creative ideas start to flow. And what I what I find interesting about creativity is that it feels very good if you nail a creative idea. Just this morning I was working on a Pretty long piece, around 2,000 words. And as a writer, you wanna make every word count, of course, but the introduction and the ending are what's absolutely most important in any good story. You can write a a a a perfect story, but if the intro is boring, no one will read it. And if you write a good story but the ending is weak, readers are going to be left with a feeling of Meh because the ending the ending should should count. And I was struggling with at the intro I had pretty nailed down quite quickly. I I had a few visions throughout the week, no I knowing that I had to sit down to write today. but the ending just didn't feel right to me. And then I went for a walk and something came to me. And I remember it just feels so good. There's this this extremely satisfying feeling once you nail the creative process. Wha why is that? Why why does creativity feel so good? I love it. so you definitely there's there's a a nice dopamine boost that you get when you get any with with any correct answer, right? We we we know we can study this in the an MRI scanner and when people get a question and and and and then you know they're they're told that it's right or they know that it's right, that their answer is right. we can see you know in these reward areas of the brain, we can see dopamine release. and so part of it is about, okay, yes, that's the right, you know, I have solved it, right? That's the right answer. And that often accompanies creativity. Now it Could accompany something like working out a math problem and figuring out what the answer is. It it accompanies creativity as well. But creativity does seem to have some uniquely beneficial and healthy elements. And they come from a few different places. One is the ownership, the feeling of efficacy and agency that you Created this thing that wouldn't have otherwise existed. Another is self-expression, depending on the kind of creativity that we're talking about, having something that you, you're trying, you're s you may be struggling with or you're trying to express well that's meaningful to you. that also is all of these things, by the way, are associated with with with dopamine transmission in in the brain as as a reward. Creativity can also be very healthy as a way of bonding people together. and those sorts of interpersonal bonds. So if you're working on a creative project with someone else, that that turns out if you co-create something, I mean, you know, to take to the extreme, if you co-create a human with someone, you're you're very bound to them. but if you co-create an idea or a product with someone, there there's there's some real sort of interpersonal. sort of oxytocin, vasopressin kinds of things that that happen with with those in interpersonal connections, all of which are related to to reward. so I would say part of it is about, you know, cracking the solution, getting it right, but part of it is about some of the things that are more specific to creativity that that that are associated with reward signals in the brain. And some of those will also be relevant in the context of what we maybe lose a bit of when we're working when we're co-creating with AI. Yeah, and I really want to get into AI also as a segue perhaps later into the AI discussion about the study that you performed on works that were created with AI. Before we move on to that, there are two questions I wanna have I wanna ask on the on the human part of things, which is creativity is a very desirable trait. We see it like w when we ask like w what do you look for in a partner? it's oftentimes people say humor, but humor requires creativity, of course. I think so. Yeah, we we think of actors and and writers and people who make music. We value those skills very high. So it's very s sought after. W why do you think that is? And do you think anyone can be creative? Because there are a lot of people out there do who say, I'm not a creative person. I I could I could never do that. And can it be practiced also? Yeah. So those are great questions. I think, you know, the evolutionary theory on this, you know, i is is is sturdy. I I wouldn't say it's sort of absolute, but there's a to me a very kind of convincing narrative about creativity as as a a human-evolved attribute that has been probably more than any other thing the reason why. You know, a creature that can't run fast or fly or has no fur and you know, no sharp teeth, not you know, the teeth are not sharp, we can't see in the dark. You know, all of these things that we can't do. We don't have a hard shell to protect us, our sense of smell is is not very good. you know, we're we're we're we're really in in many ways not very adapted for for the world around us, and yet we've we've succeeded and and sort of dominated all these other creatures that have those advantages, largely. through two things, our our capacity to, you know, create, and our capacity to collaborate, to to work together in coordinated, planful ways, right? To plan and coordinate. But invention has been a huge part of our evolutionary success. So it it makes sense you know, and it it still plays out in terms of market forces, right? If we want to sort of do the analogy between you know, evolutionary selection and success in in the market. And you can construe the market as like economic, but you can also construe it in terms of just you know, success along other dimensions, whether it's interpersonal success w or or success in perhaps a a spiritual or or other domain. those kinds of market forces tend to favor some you know you you are given credit for originating something in in many of those contexts, right? So something in in the sort of economic context has more value if it's more scarce. And and in general, novelty has been more scarce, right? This thing, or there are good ideas, but what but but you know the idea that only one person has that has more value than the idea that a lot of people have. And we can come back to that theme. But even just it turns out in intersocial dynamics, being strategic and create you know sort of inventive in how we approach those dynamics, can be a to our advantage. I will say, even to the point that, and and you talk about how creativity is a a desirable characteristic and and is generally sort of thought of in in a positive light, there's also quite a bit of work on what we call malevolent creativity, where we can succeed and in fact sometimes succeed in in sneaky or dastardly or even you know unkind ways based on creative ways of tricking other people or harming other people. And and so not every form of creativity goes to go goes toward the good, but it it does seem to be adaptive, both in terms of older, kind of longer story of of evolution, but also in in individual lives for you know, how how market forces favor creative thinking. Yeah, that makes that makes a lot of sense. And I I think it's completely right to say that part of of why we are successful and part of what it is to be human is to invent things, to solve problems and indeed do it in in a coordinated way. So I I understand why people would desire creativity in in a partner, for example. Which is why it would be interesting to know when someone is listening to this that doesn't really feel like a creative person, if you think that creativity can be trained and whether it can be instilled in in anyone. Of course, there's going to be levels, right? I mean, yeah, there's with a running, you have someone who may be a fast sprinter, there's Usain Bolt, but if you have both your legs, most people can sprint. even though it's not as fast. And I think just being creative in in a regular way might already be enough for for some people who think that they aren't creative at all. So I'm very curious to know if you think that everyone in some sense can be creative. I absolutely do. Everybody can be creative and everybody is creative probably every day. And one of the things that's been really interesting to me and and a focus for my lab's research, especially on the neuroscience of creativity, is to move away from the more traditional focus on creativity as a trait, which is to say something that is really stable and it doesn't change across a person's lifetime. You could take a test and get a number, and now that's your level of creativity, just like, okay, you take a test and this is your IQ and that's your number and go live your life. Well, what do I do with that? Right. Shift away from all of that to thinking about creativity as a state. And so specifically, just to your point, something that can change and does change over time within a life, but also within a day, within an hour, even within a minute. And looking at those neural dynamics, what is it that changes in a brain? When you go from thinking less creatively to thinking more creatively. And how can we help that happen? So the work we talked about before, where we're zapping brains with electrical stimulation, that's part of that. But also translating that into classrooms and working with schools to say how can we teach students in ways that encourage creative thinking, encourage, to your other point, creative practice, because absolutely this is the sort of thing that we can practice and get better at. In fact, What maybe the most encouraging finding in all of the creativity literature is a meta-analysis that was done. So a meta-analysis is a study of a lot of other studies to see, to see what the overall pattern is, of a lot of work that tried to help people think more creatively. Various creativity interventions. Interventions mean we we do something to change how you're behaving. and what they found across a broad range of different creativity interventions is that almost all of them worked. and I think that's a little bit of a of a of a secret that that some of us in creativity research have that like we can do a study and hey, we made people more creative. I I think that's great. But i in in in some ways it the most encouraging thing is that almost anything that you do that helps people get into a creative state. That just reminds them even, and it can be as simple. And in fact, some of the most effective interventions are just this simple as saying, please try to think creatively. And once people move into that mindset, almost everything works. Almost every way of getting them into that mindset. It might be, okay, we're gonna, we're gonna have more colorful walls, or we're gonna give you warm coffee instead of cold coffee, or you know, all of these different things. And the fact that so many of them work. Suggests that it's it's less about how you get there and more about getting there. and that, in fact, to your point, almost everyone can. and and so I I think that when we when we encounter things, like for example, my lab's done a lot of research on creativity anxiety, and anxiety specifically about thinking creatively. And what's been incredibly encouraging is to find that. This is a limitation or a restriction that people are placing on themselves. But as we talk to people about, well, look at this wonderful work that you just did when you weren't worried about it, or as we sort of trick them into thinking they're doing a math problem, but really it's something that we that that's a a creative problem. Or even if we just get them to sort of reconceptualize, really how scary is it to think creatively? And they sort of lose some of that anxiety, almost all of those things are effective. And I think for a very similar reason, that if you sometimes the blockage to getting into that creative state is a more emotional one or a more affective one or or based on anxiety. And so that that's one of probably not the entire story, but one of the examples of of how I think people who generally don't think of themselves as creative can without too much difficulty actually unlock more creativity. And to your other point, no, does that mean that y you know, and and I've used that that same analogy with you know with running and and and I think it's a good one. And again, yeah, I I I I love a good analogy, of course, but you know does it mean that I'll ever be as fast as as a as an Olympic sprinter, no matter how hard I try? No I I never will be. But I'm much but I do like to run and and I'm much faster now than I would be if I wasn't running regularly. So I've made great improvement over where I I would have otherwise been. And I think for all of us in life, that's really the goal, right? To to to be sort of the that best version of what of what you value. I completely agree. And I think one more thing to to add there, what is always so fascinating to me is the the the research. done on restrictions and creativity. Yeah. So at first glance we think the more options we have to create something, the better, right? Because well, there are more options. So we can create more things. But there's actually very good science to suggest and examples in the real world that giving yourself restrictions actually enhances your creative capabilities. I think The most famous example is the children's book writer whom I forgot the name of, but he is the author of Cat in the Hat and Yeah, Dr. Sus. Yes. And and he also wrote Green Eggs and Ham. And that book actually came to life because he made a bet with his publisher whether he was able to b that he wouldn't be able to produce a children's book. using less or exactly fifty unique English words. So he took on the challenge and he started writing the book. And because he was so limited in the words he could use, he had to be creative and use this very rhythmic, very simple pattern. And that actually produced a genius piece of work and ended up being one of the best selling children's books of all time. So I think that could also be a useful frame if you are struggling with creativity. Just yeah, creativity in general is is vague. It's too broad. But if you say, okay, where do I have to be creative? What are the things I can work with and what are the things I can't work with, that could also help you to to instill that creativity, I think. I I think that's a a perfect connection. By the way, y I don't know if you have kids, but I I I do and I've I've read those books, you know. I could probably recite them because I've read them so many times to to our kids. But yeah, I did not know that story about Dr. Seuss or real name Theodore Geisel. but that's fascinating. I think if if you don't mind, I'll steal it and use it in a in a in a class with with some students or a talk sometime. But what's so interesting about that, and what I think connects it to this thread that we're on is that overwhelming, the idea that creativity is overwhelming. And to me, one of the things, you know, any any creative act is ultimately local, right? You're acting on some set of elements that that that you're directly working with, right? The the an invention has just the parts that it has, right? a poem has just the words that it has, right? but You can get lost in the vastness of the other things that could be, right? And if you get lost there, that can counteract what we were discussing before about being truly in that, in that creative state, right? Because instead of that creative state, you're in the state of being overwhelmed or or confused or discouraged, right? But once you have those that local those local elements and you say, well, I can't wander out of here. Now I can f I I can actually direct my attention internally to a manageable set of representations. And that's again, that that once you have that internal attention on on those representations, that's where the creativity is is really gonna happen. And so getting to that state. I think can absolutely be facilitated by constraint in in the ways that that you're describing. And ultimately, any creative process has to develop those constraints because you ultimately have to be just working on some local set of things. But whether you feel like you can swim through the mass universe of all all in the vast universe of all the possibilities and make those and constrain, you know, bring create those constraints yourself. that's a much more effortful and much more daunting process. If some of those constraints are imposed, you you can get there a lot more quickly in many cases. We'll go back to the conversation in just a second, but first I have an important announcement to make. If you've made it this far, I'm going to assume that you're enjoying the show. If you want to support the show, you can do so by leaving a like or a comment if you're watching on YouTube, or by rating the episode if you're listening on Spotify or Apple Podcasts. However, the single biggest thing you can do to support the show is by sharing it with someone you think might enjoy this episode. This could be a friend, it could be a family member, it could be a colleague. It doesn't really matter. If everyone who's currently listening to this episode would share it with just one person, the show would blow up in no time. Which also means that I could go full time. That means more episodes, it means better episodes, and it also means a happy host. So Please consider sharing the episode. It really helps me a lot. And I thank you. Now let's get back to talking. Yeah. And we we've talked about the definition of creativity. We talked about why it's such a desirable trait. We talked about why it is healthy to be creative, how to unlock its potential within someone. So I think it would be we've established that creativity in humans is is very important. We for individuals it is important and as a society we think it is something to to strive for. However, we see the rise of artificial intelligence, which puts the whole creativity discussion in new light and and makes it a lot more difficult. I think we have to re-establish what it means to be creative and and what a creative work should consist of. And before we talk about your your study that you've performed on on the use of of Chat GPT, do you think AI can can truly be creative in the sense how we've just established it? So I I think there's no question that AI can generate products that meet the product definition of creativity as novel and useful. The process side is one that I'm more invested in and more interested in. And I think it's very informative. And as a neuroscientist, one of the things that's been fascinating for me with the sort of rapid expansion of of AI in creative contexts in the last few years has been to really consider the differences in process for a human brain and a a language model. Most I'm mostly we're talking about large, you know, large language models when we have these conversations. Although a diffusion model in many ways is working you know, in you can make a lot of sort of analogous conclusions. But so so whether we consider that process to be creative, I think you know, this process definition that that we developed for creativity is very much a human process definition. But at the time that we were doing most of the thinking about it, that was the only game in town, right? That was the only kind of creative process that that really needed to be considered. We we anticipated some in that paper, you know, how might this play out in in AI and and you know, we'll have to wait and see. I think since then we know a lot more. I am quite comfortable saying that what a language model does is a form. of a creative process. I actually don't think that's too you know that that that's it's it I I I don't I I'm not sort of I don't begrudge and you know like I I I I I think some people don't want that to be true and and I understand that. But I think as a you know I I I think our job as as researchers, as scientists is is to take as as objective a look as we can, you know, at these things. And I I think it's entirely fair to Call what AI is doing a kind of creative process. But I think the key thing is to understand the differences between that process and the human process and what that means about the kinds of creativity and the landscape of creativity that we are going to be now interacting with. And especially how though. Those differences, if we're not careful, can start to erode some of what is most valuable about human creativity. Yeah, I I agree. And what what's so difficult for me about this whole creativity debate in context of artificial intelligence is that first of all, I have to do creative things for my work. I I write a lot of articles, I have a lot of conversations. with people. I I used to make a lot of videos. So I but I also use a lot of AI in my in my day-to-day life. I think AI is great. I think it's extremely helpful. I'll give you an example. This interview that I did, it it lasted over two hours and I had to condense that into 2,000 words. Okay. What I used to have to do was I would have to listen back to the entire interview, either typing away or with a notepad, which had some benefits because then I could improve on my interviewing skills. But it wasn't the most fun part of the job, right? I think writing is fun. I think doing the interview is fun. Listen listening back to the interview, not so fun. So what I do is I just run the transcript, I turn it into a transcript, which is AI, and then I throw the transcript into Claude, and then I say, hey, categorize all the things he or she said into usable text and and don't paraphrase, just use it exactly as this person said it. And now as I'm writing, I can just go back and like, w what what did he say about the beginning of the company again? Let me just check and then I immediately have it there. So it is very helpful in my creative process. Also, I got surgery in January this year, knee surgery, and I wasn't able to work for for two months. And I, for the first time, had some time on my hands, at least for the first time in a long time. And I started looking into vibe coding, essentially building websites and building games with AI. And I used to have all these ideas for apps that I wanted to build, but I didn't have the skill to code, and I didn't have the money to invest in in a coder. And now all of a sudden, from from my bed with my leg up, I could just turn these ideas into existence. And to me, that was just so such a good feeling that I was finally able to to to turn my creative ideas, if you will. To life. And now my coders might hate me for this, but I think that when a person wants to go on a website or or perhaps play a game on their phone, they don't generally care how that code came to life. However, when we are reading a book, for example, we care a lot whether the the the author has used AI to make that piece. So I I do think and I'm very biased here, but that there is a a big distinction be b between how we view those end products, if you will. yeah. The thing that you looked into were I believe there were papers published by by students before and after ChatGPT came to market, at least products produced by by students and And you you've looked into whether they yeah, what the effect was essentially. And I think, and again, correct me if I'm wrong, that the finding was on the surface it looks more creative because there was a wider variety in the words used. However, the ideas seemed to converge towards each other. So the actual creative substance of those works were less original. they only looked more original because the use of of fancy words. Which I think is i very dangerous, if you frankly, if you if you ask me. So all of this is to say what what did you do exactly during that study? But also are you worried at all? How d are we going to lose our creativity if we continue to use like this? Because we can use it for creative things, but we can also Do it to yeah, remove the creativity from our lives. Well, I'll start with what your story sort of sparked. and I'll go on to to our study. And and and so the thing that I really liked about both of the examples you gave. So where you're having Claude categorize things from the interview or where you're vibe coding with your ideas. You have the story idea, and you're gonna write the story based on the narrative that you are creating. and to sort of generate or or reveal the insights that you want to reveal, that's that's your idea that you're starting with. W when you are vibe coding, you're having, you know, the the actual coding is not done by you, but the idea. Is yours. I think the one most important thing for us to do when we work with AI, if we want to retain our creative value and our creative experience, is to start with our own ideas. That's the single most important and and you know, the the other side of that potentially dangerous thing about working with AI. If if we start with the AI ideas, That's where this phenomenon that we've been observing can really start to manifest. And so in the study that you're talking about and in other projects we've done, we've looked, we're looking at human writing in in in some of this work on really large scales, hundreds of thousands in this case of essays that students in the US write when they're applying to colleges. And these are stories, these are these are essays about their own lives. They're supposed to be sort of unique ways of telling their own story, right? And what makes them who they are. And and as you say, what we found is that right around 2022, going into 2023, magically everybody's life story became more similar. and it but the the really scary thing, which again you you point to is The surface level actually became more creative in appearance, even as the underlying ideas converged more. And the way that when I say that the surface looked more creative, yes, the words are more flowery, but it's not so much about those kind of fancy words as it is about using words that have these different meanings. Coming back to your first question about semantic distance and how we studied it even in the early, you know, the early days of doing this, these essays involve words that have meanings that are diff far from each other. And so the ways that we assess creativity say, okay, that's creative writing. That's a creative essay. But it turns out that because of the way that LLMs function differently from human brains when they create, LLMs can tell the same story with a great variety of words, but they're still telling that same story or small variations on that story over and over again. When a human brain is generating ideas and when a human brain is generating words, it's using a shared network, a network of semantics, meanings that that person has developed over the course of their own unique experience. What that means is that it is much more difficult for us to separate the sort of word originality from the idea originality. And in an LLM, those two things can be much more easily separated. An LLM can sit in the same neighborhood of idea space and pull in words from all over that semantic. embedding space that we started out by talking about. And so that can it can a human brain do that? Yes. But it is much harder. It is much th those two I levels of words and ideas are much more connected in human brains because of how we work. we'd have to really try if we wanted to separate those. But what we found is that when humans wrote essays, those two levels, the creativity at the word level and at the idea level, they were connected. They they were associated, positively associated. When the LLMs are writing the stories, those two levels not only become more weakly associated, in many cases they actually reverse that association. So what that means is that the words that seem the most creative on the surface are actually telling the stories that are the least original in terms of the ideas. And that to me is particularly scary as you suggest because Since ideas and words have always been linked throughout human history, when other humans read our work, they tend to give us higher marks for creativity based on that word level. So even creativity experts in our study were tricked. They didn't identify that these were that these different essays had become more similar. They thought they were more creative because they were getting more. variegated at that suit at that superficial word level. What that means is that we have to rethink how we assess creativity. We can't trust our own instincts anymore. We need to use tools that allow us to map this sameness of ideas. So yes, humans should do the reading and form an impression, but they should also be informed by a map of idea space. And that's something that we've been working on. And in fact One of our newest findings is that you can map these idea spaces and you can find AI's blind spots where AI doesn't produce ideas. And you can look at the human ideas in those blind spots. And what we found is that the human ideas in those blind spots are just as good as the AI ideas, and as the human ideas that look more like AI ideas. That's really important to us. Because what it means is that when AI homogenizes our ideas, it's not just getting rid of the bad stuff. It's getting rid of good stuff. And if we have less variety of good ideas, we are much less likely to solve the problems that really need to be solved. True, like the likelihood of innovation over any smaller or larger context depends largely on the number and variety of viable ideas. So it's not that AI ideas are bad, it's that human ideas are good. And we lose them when we get sucked into AI homogenization. But and I know we're we're almost out of time, but I would just say the most powerful thing that any of us can do to keep that from happening is to start with our own ideas. And we're working now with schools to help students do that. One of the things we're doing to tie these different threads together that we've talked about, Julius, is actually give kids neurofeedback on when their attention is. Too much focused on what the LLM is telling them and not internally focused enough to help them keep that balance so that their attention can be focused on their own thinking, even while they work with AI. Yeah, the the one thing I I advise everyone when working with AI, if they still want to learn something and actually take something away from their session with Chat GPT or Claude or whatever, is something That is called Socratic learning. So by doing this, you basically tell the AI, you ask me questions and let me formulate my words and come to my own conclusions. And then where I get off the rails or whether I have something completely wrong, you can steer, of course. But this is a good way. where you do still do the thinking because it's just that I see people on on LinkedIn, I see people on Twitter, and and they they're they're even using AI to produce their own comments or or or share something highly personal. And nobody logs into Twitter or or or LinkedIn or or whichever social media platform to read like super polished, fancy words. No, we want character and we love those less polished views. I think that's that's much better than than when everyone just st sounds starts to sound the same. So for me at least that's a good approach if I actually want to walk away with some knowledge. Just let me do the talking, just ask me the questions because it's it forces forces me to think. I think before we before we end it today, I I saw a c a quote of yours If your ideas are just AI ideas, you're replaceable. The future belongs to those who still think for themselves. I think that's a really important message, especially for younger people, because they in our work with younger people, often what we hear is that they're very discouraged and that, you know, what what is the point of trying to learn in school? What is the point of trying to improve at anything when there's this massive intelligence that's always going to be better than me at everything? It comes back to something we brought up before. What we're finding about human creativity in the AI era is that it's not about having a better idea, it's about having a more distinct idea. And so AI ideas are good, human ideas are good, but everybody shares these homogenized AI ideas. And it's the people who think for themselves who will have the more valuable ideas, because as we talked about, A good idea that everybody has is not very valuable. A good idea that no one else has is extremely valuable. And so when we show students that their ideas can grow outside of these homogenization clusters that AI tries to suck us into. And when we show them that they got there by thinking for themselves, it's like showing them a superpower. And we can make this very real, very legitimate claim about why that kind of thinking is only going to have more value. Because most people will tend to do the easier thing. And if you're someone who continues to think for yourself, you will have more value even than you had in the pre AI era. That's a beautiful way to end it. Adam, thank you so much for for this this conversation today. this was Phenomenal. I had a great time. You teased us with a paper that was coming our way in the beginning of this episode. Yeah. Where where can people go if they want to keep up with everything you're doing? Maybe look up some of your previous work and also be on the lookout for for new stuff coming out. Yeah, well, so you can certainly go to our lab's website and we post those you know those publications. I I would also go to the Society for the Neuroscience of Creativity. SFNC, which is a great kind of hub for all this kind of work happening in all of the different sectors that we talked about from neuroscience to cognition, a lot happening in AI, a lot happening in health and flourishing. We are also this summer in in one in London and one in Oxford, we're gonna have the the an SFNC meeting on w which will touch on a lot of these themes and then a conference in Oxford funded by the Templeton Foundation that I'm coordinating specifically on AI and human flourishing. and that'll be June twenty first and twenty second at Oxford. Wow can can it be attended or is it Absolutely. Yes. Awesome. Well I'll put that in my calendar for that is not quite up yet, but it will be soon. Yeah. Okay. Well I'll put that in my calendar because that sounds something I'd love to to be at. We encourage you and and what I love about SFNC, which we we do that meeting every year is is that it does bring together people not just who are academics but a lot of artists and a lot of people in media. I think it it's great actually to have and people in industry as well. it's great to have that sort of melting pot. I agree. I agree. Well I'll I'll be on the lookout for a bunch of things now. Thank you thank you so much Adam. This was awesome. And Yeah, I think the last discussions about AI and and creativity, d this will be very relevant for for a very long time. So yeah, perhaps we'll run it back someday in the future to do an episode specifically about that topic because this was great. Well thanks, Julius. I I'd I'd I'd welcome that and I really enjoyed our conversation. Thanks for having me.