Diverse Intelligences in Natural, Artificial and Hybrid Systems - A Conversation with Prof. Sheila Macrine and Dr. Jennifer Fugate

Welcome to the Future Learning Design Podcast.

I think that this pluralistic framework that we are trying to advance in terms of an embodied, situated, emergent, broadening the view of intelligence, it really opens up the gate in terms of like, what do we mean by this? And what is the purpose for these intelligence tests?

Hi everyone, and welcome back to the podcast. My name is Tim Logan, and the podcast is produced by Good Impact Labs. We all take lots of different roles in our personal and professional lives, but one of the roles that I find most exciting and aspire to myself is being a convener and a connector, bringing people together for generative purposes and in ways that don't always happen in the silos of modern institutions. My guests this week, Professors Sheila McRae and Jennifer Fugate, have been doing just that in the emerging field of diverse intelligences. They are gathering some of the most leading-edge and brilliant researchers working today on the problems of how to understand agency, cognition, and learning in the diverse ways that they are showing up in all kinds of bodyminds— natural, artificial, hybrid, and everything in between. It's an incredibly exciting area of research that is challenging us all, but particularly the legacy systems, to loosen its grip on outdated and limiting ideas like IQ. Sheila and Jennifer have amazing and diverse expertise from a variety of disciplines, so I'll let them introduce themselves in a moment. But just to highlight, the 2 particular places where they have brought these researchers and ideas together are in 2 co-edited volumes, both with MIT Press and both open access. In 2020, they published Movement Matters: How Embodied Cognition Informs Teaching and Learning. And in June this year, they released their second volume entitled Embodied Intelligence: Multidisciplinary Perspectives on Natural, Artificial, and Hybrid Systems. You can find links to both in the show notes. I'm really curious how you both see your role in the space, like given how diverse your backgrounds are and all the different things that you do in all sorts of different spaces, education and adjacent, you've kind of ended up in this really interesting space of convening all these incredible people, including yourselves, of course.

I know.

In these books, and they're, you know, they feel like really important books. I mean, not, not just to, you know, fluff your egos at this point, but they do feel like really significant books of bringing together many of the prominent eminent thinkers in this space, which feels like an, to me, an incredibly alive space of new insights. I mean, you know, there's so much in there, but yeah, so perhaps we could start there. How do you see your role? Did you just kind of somehow accidentally fall into that space or was it intentional? Where, how did that happen for you?

So I am trained as a cognitive psychologist and for years my area of expertise was intelligence, human intelligence. And I had a lot of experience with clinical assessment of children with learning disabilities, autism across the board in hospital settings and in private practice and in schools. And I got hired as a professor at a university, not UMass, before I came to UMass and I was directing the special education program for teachers. My graduate studies involved a lot of phenomenology, and that's how I kind of came into embodiment. And I think in my head, I feel like sort of a philosopher and a phenomenologist along with having a psychometric background. And that was really frictional for me in my career because, you know, universities really have their disciplinary silos. And so doing research in one area and then trying to do the psychometrician stuff that I was trained to do, there was like a clash and I was trying to justify like, how do these two fit together? And of course, you know, eventually it got to the point where I started really writing more about embodied cognition in the early 2000s. And then I came to UMass in 2011. And I think it was around the time Jennifer came, but we really didn't click or meet and talk until maybe 2015. And we started talking and we both had this real interest in embodied cognition. And that started us on this whole thing in terms of publishing in a lot of different venues in terms of just trying to get it out there. And I guess my motivation was to help teachers understand the way that humans really learn rather than having teachers transmit information, but to really like kind of look at embodied cognition in terms of mind, body, environment, and how that environment can impact And all the things that we're going to talk about in terms of how teachers can implement this. So as we started writing, you know, I started thinking about like, how can we translate this for teachers? Because I don't know how it is in Europe, but in the United States, we have a Department of Education and all the research that's being done doesn't really trickle down to the classroom for like 5 to 10 years. And my idea was that if we could get a book out there that would show teachers how to do it that would be open access, that would be accessible. And we made the pitch to MIT and they really were excited about it. So that book was published in 2022. By the time 2023 came, we were seeing embodiment exploding all over the place. Everybody from all the disciplines were adopting the term and it really bothered me. And I told Jennifer, I said, these people are using embodiment wrong. They don't have a precise understanding of what it means. You know, like if you have a robot with a head, then it's embodied. So I think that's where our motivation came, that we would have a common language across disciplines. And so we made the pitch to MIT, and I was thinking that we should contrast human intelligence with artificial intelligence. And we submitted that. And they came back to us after their editorial board reviewed it and said we were too human-centric, that intelligence is not just human. It goes all the way down. And so that led us into a deep dive into what's out there. And quite frankly, I have to tell you the generosity of Carl Friston, you know, just as a cold call for me. Andy Clark, he endorsed the first book. He was really cool. And, you know, Rolf Pfeiffer. And then that kind of led us down the road to people like Michael Levin and Josh Baumgart. So I feel like there's a certain kind of zeitgeist going on right now in the world that there is this movement for cross-disciplinary cross-pollination of all of these different people in, you know, different disciplines trying to break down those silos. Now, the universities don't do that, but I think there is a group of people now that it's just a very exciting time.

I love it. I feel the same. I completely agree. It's interesting, isn't it, how these liminal spaces like podcasts or these YouTube exchanges or all these different spaces and, you know, these communities cropping up where it is happening, even if it's not happening in the main institutional spaces. Right.

Exactly.

Jennifer, did you want to add anything to that?

Yeah.

So much like Sheila, I think we kind of came about this in different ways. So I've been trained actually as a comparative psychologist. I got my degree working in animal behavior, but I've always been interested in evolution of the brain, evolution of cognition. And so I had the pleasure of actually working with Larry Barceló when I was at Emory as a graduate student. And he was just, just foundational in my knowledge of embodied cognition. And interestingly enough, though, I went on and I did a postdoc more with human cognition, specifically in emotion. And the whole time, though, that I was working in emotion, which is my other field, I was really still drawn to this idea of embodied cognition. But I didn't have— I don't think I had the space yet. This was almost 20 years ago where people were really talking about it in the way that I wanted to see it represented. And so after, when I joined the faculty at UMass, where I met Sheila, and we started talking about it and started realizing this read and the philosophical implications and how it draws from psychology and cognitive science. And as she said, all of these people are using this word but in a very different way. And so I think that's when we really started to think about, well, what is the bigger picture here? What is the message? How do we get rid of these different silos? And how do we really engage academia as a whole to talk about this really important concept? And, you know, I can't extend my gratitude enough to Sheila for allowing me to write with her and be a part of these last 2 volumes. So now I actually work in a clinical psychology program, although, you know, I'm not a clinician. And I think that's been a really interesting piece of the puzzle as well, because we're starting to think about embodied cognition for therapy as well. And what does that mean in terms of movement of body and interoception and mindfulness and how that can regulate emotion? And that's the other kind of side of things that I do in my research lab.

Fascinating. Wow. Interesting. I mean, it's just so interesting to me, like regular listeners will know, I've done a lot of over the last 2, 3 years kind of dives into this space. I'm not, you know, nowhere near my own specialism. I, you know, trained in anthropology and much more interested in the kind of social and cultural aspects, but just opening up that space of what is it? How do organisms actually learn? And so, you know, ecological psychology, active inference, embodiment, radical embodiment, all of the good stuff that you engage with all the time. I've just been on this journey of learning myself and then engaging with people like Tony Camaro and Dore Abramson and quite a lot of the contributors to your books and Joe Bowler and just great, great stuff. And so I share your feeling of this kind of excitement and zeitgeistness.

Yes.

A danger of that becoming overused and lose some of the integrity or whatever, but At the same time, I think a lot of teachers feeling like there's something doesn't quite match the reality that I see in, in these wriggling organisms that I'm, you know, caring for in these classes, right? It's the things that I've been told about how learning happens, cognitive load theory, or all these kind of more cognitivist ideas just don't necessarily match the emotional and relational realities. And so that's where it's coming from for me in terms of an inquiry, I suppose. And I'll put definitely links to the books in. I love the fact that they're open access. I think that's so good.

Oh, thank you to MIT Press for that.

Yeah, for, yeah, no, absolutely. But I think so important. I mean, so often these academic books are just prohibitively expensive, right? But that's just amazing. So I'll put links to the Movement Matters book in the show notes, but particularly I would love if we can focus the conversation around the new book. Super exciting. And it's interesting to me to hear about this, this controversial word intelligence that shows up a lot. in schools and, you know, in people's lives and your reflections about psychometrics, Sheila, and the friction. It's, you know, there's some tensions there for me that I'd love to get into, but the, it's really interesting in the foreword to the book. I don't know whether I pronounce his name right, but Aida calls it, what does he call it? A bold manifesto for a paradigm shift in how we understand intelligence.

Right.

Yes to that. That is like, I mean, don't we need that? And so maybe, I mean, that's a kind of a good place to start. How have you approached that? I mean, this, bearing in mind what you've just said, Sheila, this is probably like your entire career's worth of work I'm asking you to summarize in a few sentences. But to me, there are lots of tensions in the way that intelligence has been thought about. Galton and eugenics and, you know, all of the dark history of the concept of intelligence. But it's also such a present concept in educational spaces. So just opening that up, where would you take that in terms of the inspiration for the book and what you feel perhaps the book is starting to lay out in terms of a new understanding?

Right. So my training as a psychometrician, specifically intelligence, I think that the way that you asked the question about how it was weaponized, I think what happens is in Western society, you know, intelligence was institutional education, psychometrics, a monolith, brain-bound. You know, I know as a teacher training, you know all about social cognition by Vygotsky, Piaget, and everything. And as much as we talked about transformative education for teachers, we still had them in little rows and columns, and it was still a transmission. We talked about this sort of interactive teaching, and, you know, we've all had that philosophy in education. But the reality was, is that I think what we're trying to promote is that from, you know, our earlier work in terms of movement, in terms of the whole person, the whole, you know, looping of the mind, body, and environment, and how can teachers take advantage of that. So intelligence was really a reductionist model. And, you know, we always read in our psychology books about the brain in a vat. And I think that the educational pedagogy that we were all exposed to really sees the student as this receptor of the information that the teacher or whoever is transmitting that information. And I think that it was a trap. And I think it was like sort of a a sheep in wolf's clothing, sort of, that we were selling this to teachers. This is gonna work and this is gonna work and all these professional developments, this is gonna work. And the reality is, is that we're really kind of replicating what was done from the beginning of education. It's really transmission. So I think that the reconceptualization of an embodied intelligence means that we really have to think about it in terms of multidimensional, that we can't just put kids in little blocks. And I was guilty too of diagnosing kids with a specific learning disability, and this is the part of the brain that's functioning, or, you know, hemispherical dominance and all the other stuff that goes into psychometricians' toolbook. So I think that what Foumya was trying to say in terms of the computer metaphor of intelligence or the brain is no longer applicable. It's not a software running on a computer, and that this feedback loop, we really need to take advantage of it. So that's just not the student, but it's also the teacher and this kind of dance in terms of understanding and extending their learning experience in terms of mind-body environment. And then the environment that we propose, you know, how we can best help learners is to, you know, recreate it in terms of having the teacher really use more physicality, more gestures, more dynamic intervention with the students, and then having the students sort of, you know, mimic or copy that. This is how we think learning would best do. But I think that, that what Fumio was trying to say is that, you know, it's beyond the biological-artificial divide that we really have to think of intelligence as something greater than just this idea that we have of this brain in the vat. Do you want to say something about that, Jennifer?

Yeah, I think I would just extend, you know, there's this idea, and obviously I've been guilty of it too, and many of us instructors are, where you teach the concept, it's like formalisms first, and then if there's time, Let's do an example. Let's walk through an example. Well, what if we flipped that model and started with the example and then abstracted the knowledge from that? And I use abstracted in a different sense of the word, not actually meaning abstracted in an unembodied sense, but what if we learned through our interactions and then we'll be able to take the principles from that learning? And here I think is where Sheila and I really resonate with one another. This idea of this interaction with the teacher, the teacher being in the learning. You know, this goes back all the way through to Merleau-Ponty and Montessori and all of the other philosophers in early education. Even Dewey talked about this when we talk about learning and then how we get to more formalized education. And of course there's, you know, there's controversy in how that happens, whether at some level thought is actually abstracted from the experiences or simulated. But this idea, I think nonetheless, regardless of the mechanism, that we need to start with the experience and then look at the resulting knowledge. And I think that really kind of flips the model on its head.

Great. Yeah, no, I love it. And it's interesting to me that whether it's flipped or not flipped, where the value gets put is also an interesting question, right? Because however you get there, The thing that often we value is the abstracted, disembedded kind of intellectualizing of the concept rather than the concrete experience of being or doing. And it's interesting how deep those biases are. I mean, you know, we are all testimonialists for Guy and Emily's book, and that's, you know, that's part of what they're saying there, right? Is that it's just, it's so deep in the psyche of the formal education system that However we get there, the thing that we really value is getting there to this abstracted conceptual idea or this kind of abstracted, disembedded intelligence concept, which I think Guy even talks about as not really even having a basis in science, but just intelligence is what the intelligence tests measure. So you end up with this weird circularity of what is it actually connecting with in the real world?

Sure. And that's a problem for assessment too, that goes right along with it. You know, we're assessing students in a very unembodied way in the sense, you know, with our formalisms first and our, you know, multiple choice questions and things that really resonate more with a disembodied view of how learning works. And so I think thinking about how we assess in ways that are much more embodied is just as important. And just by way of example, one of the things that's really been unique in the department I work at now is looking at how clinical psychology students learn through their simulated experience. So One of the nice things about the university I'm at is that students in multiple classes throughout their development are assessed on their behavioral competencies for internship through simulated learning. So actually working with simulated patients in our center so that they actually get that practice. So I think it's a great thing that our program does. And part of my research there is to track that learning and how that works through simulated experiences.

Fascinating.

I think, uh, for me, testing kids with the IQ test, that's probably as archaic as the Model T Ford. We haven't changed it much since the 1930s. You know, this is a very biased, culturally biased test, and it only taps into certain things that they hold important in school. And I think we're moving away from sort of this elitist brain-centric hierarchy of intelligence. And, you know, it's almost a gatekeeper. of certain class systems and everything. It's like really, like you mentioned eugenics and everything. So I think that this pluralistic framework that we are trying to advance in terms of an embodied, situated, emergent, broadening the view of intelligence and recognize what is an intelligent agent. And that's kind of like where I'm at right now. I'm just kind of doing this deep dive in terms of what is intelligence? You know, we talk about human intelligence, we talk about animal intelligence, we talk about birds that plan, you know, like what, how do we define and intelligent agent. And then, you know, even with Michael Levin's work, he's going all the way down to the cellular and molecular level. He's talking about that if it's a self-maintaining organism and it has a boundary, you know, the Markov blanket and everything, that it is an entity and that it has a certain amount of intelligence because it's seeking food for sustenance, for self-maintaining in existence. So when you look at sort of the notion of an embodied intelligence on a spectrum, It really opens up the gate, you know, in terms of like, what do we mean by this? And what is the purpose for these intelligence tests? And I have this one set of ideas in my head, and then I teach a very sort of traditional model. And, you know, I have this argument with other people and, you know, they say, how can you teach that when you believe something else? You write about something else. And I say, but we have to arm teachers with the best qualities or the best skills to assess kids to get federal money, to get state money for special education for kids that need extra help and stuff. So that model still exists. And so as much as I revolt against this traditional model of, you know, brain in a vat and these kind of sorting systems that we have for children, that in my ethical and moral soul, I don't believe in it, but I have to kind of prepare teachers for what the reality is in school. So it's a real dichotomy for me, and I'm still struggling with it. And I talk to Jennifer all the time about it. You know, I'm teaching them how to give IQ tests, how to give achievement tests. I'm teaching a traditional psychometric. That's my job. And so that's where the conflict comes in for me, both personally and professionally. I think that the university, in terms of what my scholarship says, and then what I have to implement in terms of teaching teachers who work with special needs kids who learn differently, that we have to really kind of influence the educational system. And we're not there. We're not even close to that. So when I'm looking in one place about this gestalt of what we're doing in terms of what intelligence is, of what cognition is, of what it means to, you know, understand our environment, we're not doing that in schools at all.

Yeah. No, I mean, I really appreciate you naming that because I think that dissonance is a reality for so many people, right? There is, as you said, it's like there are gatekeeper concepts and gatekeeper approaches, which are the allocation of resources or are the allocation of federal funds, like you say, or, and you, to an extent, you have to be pragmatic and play that game. Otherwise you might end up not supporting the, you know, the marginalized young people or whoever it is that you're working with, any young person who needs support and deserves the best possible situation.

Yeah.

And at the same time, almost like living this double life, right? I mean, I feel a little bit the same. You kind of, And you do whatever you can, publishing books, sharing podcasts, et cetera, to kind of influence the conversation and move the needle to say, actually, the science is coming. It's there, actually. It like, it really is. I mean, so much of this, the researchers' work that you share in the book is so exciting, I think. And so like, just absolutely cutting-edge stuff. And then we go back to work and we go back to school, right? I mean, it's just, it's such a dissonance.

Yeah.

So I appreciate you saying it, but perhaps you could say just, Jennifer, maybe just say a little bit more about that kind of, when we talk about plural or like Michael talks about diverse intelligences, it's quite hard for people to wrap their head around what that might even mean. And until you start hearing some of his examples, it's like, it's wild, right?

Sure.

But it, because we're so anchored in a particular dogma around these concepts, it's quite difficult to even imagine what diverse intelligences might mean.

Yeah. And as an evolutionary psychologist, of course, that resonates a lot with me. I just want to jump back to Sheila's point for one moment if I can before I answer that question. So I think, you know, policy is always chasing the science and that, that's not unusual in this field or any other field. You know, I think about even where we are with AI in terms of regulations. You know, we have just pushed through so much and the regulations aren't there. So now People are having this watershed moment of, oh boy, we better have policies, we better have safeguards, we better put guardrails on this. But that's the thing that happens 10 years down the road, right? So in some ways, I don't think that the education system and the ability to catch up with the science is unique in how slow it is. But I do think it's something that we need to think about. To go back to your other question, though. So, you know, as a comparative evolutionary psychologist, I think about these things a lot, and you always think about what is evolutionarily adaptive for an organism. And so, you know, we have goals that are proximate. We have also, you know, ultimate goals. But in terms of when you're thinking about animal cognition or even single-cell cognition like Michael Levin does, and thinking about this cognitive light cone, it's like, what is adaptive in the moment given the context, given the organism's Umfeld? And so, you know, that can look very different for different organisms, but it can also be— I think what we're trying to get at in the book is it's determined on the body itself of the organism. So something that has a design that's single-celled or, you know, works in a way that we as humans maybe don't communicate, it could be, you know, electricity or any other domains that we don't particularly see in Human cognition, you have to think about it as what was adaptive for that organism in that point of time. And thinking about things like goal-directed behavior, I think that you can look at that at multiple levels. And I think that when we think about that, we often think about large goals and we think about human-centric goals. And I think that's also part of the book is to break down this anthropomorphism. But goal-centered behavior is really what's functional for the organism in that moment. in terms of its reproductive survival. So I think it's just a little bit of a shift and goes back to the point, I think, and MIT was right on to push back and say, we need to get rid of this human-centric model. And you kind of need to think about cognition and intelligence across evolution. And I think that was a really profound addition that we made to the book before it was released.

So cool. So cool. It's just so exciting to me. And when you, I was watching a One of Michael's latest talks the other day. And it's wild when you, in terms of the genuine inclusivity of being open to recognizing things that are not like us, right? As humans, right? Like the, the fact that maybe a, a sorting algorithm could show some kind of goal-directed behavior that's kind of somehow autonomous and that wasn't coded in by the— I mean, it's, it's wild what he's kind of allowing into the experimental space to say. Let's just have a look and see what happens rather than just define the fact that of course it's not intelligent. Of course there's no cognition there because it's just a piece of code or it's just a, you know, it's just a single cell. I mean, it's just, it's fascinating.

So I'm working with Michael on a paper, a new paper I'm really struggling with, but it's all about agency and the difference between, you know, what is an agent? What do we consider an agent? And it's very exciting. And he's opened up my eyes to a lot of different ways of formulating that question. So like, if we're talking about, you know, a human agent versus a cellular agent versus an LLM agent, like how do we differentiate those goals? You know, achieving the goals, is it, you know, endogenous or exogenous? Like who drives the bus, in other words? And I use that metaphor about like the engine being sort of the cognitive capacity and then the driver being the goal-seeking behavior. And then the interface being the embodiment. So I think that once I get that sorted out, I'll be able to publish that. And I have just submitted a paper to Artificial General Intelligence 26, and they accepted it, and it was called Beyond the Engine. So what I'm suggesting is that we have to look beyond the intelligent or the cognitive capacity and look at what the mechanisms are that create the agent. So it's not just the cognitive part, you know, when we're getting down to cellular level, or if we're looking at sort of these hybrid living robots like xenobots and anthropots, you know, like how do we determine what is an agent? You know, it's not just human agency we're talking about. So I'm really excited about that next iteration.

All right, cool. It's so interesting because like going through my mind now is like, why does this matter to a teacher in a classroom, right? And it's so interesting because I sometimes wonder like, is this just some weird curiosity that I have of diving down these rabbit holes and because I'm just interested, but I genuinely, this wasn't on my list of questions, but if you don't mind, this morning I was thinking about rigor, right? And so often in a classroom, you know, there's this kind of academic rigor or it's like we try to bring this thing back to this rigorous process of really helping our students to kind of, to get to know the world in a very deep way. And we kind of, we give it these names like rigor. But I was thinking something like, surely this is a very unformed thought, so I apologize in advance, but is it not much more rigorous, inverted commas, to acknowledge the interplay between the things you know and sense, your environment, what your body is doing, what other people are doing? There's a sense to me that that's like so much more rigorous when you open up that space than Just focusing in on, can you frame a particular proposition about this concept? Like, can you say it smartly? Right? And it feels like these are just nice to haves, like as you were talking about, Jennifer, like these experiences or these simu— you know, the simulations of the, not the formalisms. I don't know if, can we challenge that idea that actually the rigorous thing to do is this thing we're talking about, which is genuinely opening up the space of The interaction between what you care about, the context, what your body is doing and sensing, all of these things. I don't know. I don't know how any of that lands with you.

I see what you're saying, and I think it makes me think about interoception. I know I used that word before, but we know that the way that you sense your body, it's not just a skill, but it's something that can be improved and its connection with mental health and its connection Even with, I think, learning to a large extent. So when you're thinking about rigor, I think about it almost as one's ability to be in tune with oneself and how oneself is interacting with the environment. And to that extent, I think interoception's a big piece that hasn't been studied to its full capacity yet. But I think the ability to take that sensory information and use it in a way that is fluid and kind of drives Conceptualization and concepts in a way that is probably not as maybe pre-described as we might think so. But that fluid interaction, I think, between the body sensations itself and the concepts which ultimately are constituted from those experiences, I think maybe that's what I'm hearing you say when you say this type of remembering and how it feels very different, literally feels different than, you know, memorizing some kind of, you know, statement or proposition. Propositional knowledge.

Yeah, it's like the map and the territory, right? It's like you're just staying in the map. Let's learn the map, but we're not really allowing an exploration into the territory.

Yeah, just thinking, you know, one of the things that we were going to talk about too is how does this relate to large language models and AI? And I think maybe there's a connection here as well in the sense that we know these models are rigorous in the sense that they're very powerful and they're, you know, can see patterns that humans can't in the way they represent knowledge. But what I think I would push back on is that they're still really accessing text as kind of a linguistic trace. They're not actually going through the movement itself. And so we do have things like embodied robotics now where, you know, we do have systems that can interact with their environment. So I think large language models for the most part are still stuck in this place where they're learning and they're learning very effectively but they're not learning through the embodied way. And I think it gives us this almost like as-if proposition, like it feels like they're very real and they feel like they're very human. And Sheila's written about this in her recent Hijack paper. But this isn't a different argument than we have in comparative cognition where, you know, we say animals function as if they have these cognitive concepts. But are they actually functioning in the way that a human or an an embodied organism would function. So are they accessing just the traces or are they actually accessing the experiences? And I think that's where things like embodied robotics and things where robots and systems that are actually moving and sensing their space are learning fundamentally different than something that's just propagating and searching text.

It really kind of opens up so many ideas. I mean, we've just written about embodiment in therapy and, you know, having this meta-awareness with the client and everything. But now that Jennifer's brought up the AI stuff.

No, let's go there. Let's go there. 'Cause I think that would be good. One of the things I did pick out in my research, which was in your Embodied Hijack paper, that phrase, I really liked it. It was almost a kind of secondary phrase, but children construct theory of mind through embodied friction with biological agents. And that's some of, I feel like, where we're dancing around in this part of the conversation that, as you said, Jennifer, the LLMs just don't do. They don't engage with that at all. Yeah, Sheila, what are we learning about human intelligence?

I started thinking about the LLMs, and they're not the LLMs, you know, from Geoffrey Hinton's early work in terms of an LLM having lots and lots of words and reading everything and just, you know, predicting what word comes next. Now the LLMs look at vectors and similarities, and, you know, it's really, really unbelievable what they're capable of doing. And, you know, in terms of like agency and, you know, who's driving the bus and stuff like that, I just started thinking about like, have you ever said thank you to an AI system? And, you know, and I do it all the time. Everybody does it.

Yeah.

They're anthropomorphizing. And so I started thinking about it and, you know, I'm really happy with that paper. I think what it is is sort of a naive anthropomorphism. And in terms of children and schooling and everything else, I think that the point that I was trying to make in there is that just like Jennifer is talking about this evolutionary cycle that humans went through, that we respond to living entities, whether it's, you know, another person or if it's an animal in distress or, you know, any kind of signal that we get from a living biological being. And our brains are kind of formed to respond to that. Then when you have an LLM that is really kind of engaging with you in a way that seems like a social cognition. It's very deceptive. And what I'm suggesting is that it's hijacking our emotions and hijacking our embodiment. And I think that this has totally been undertheorized. And I think that in terms of what the responsible thing to do is to have the people that develop these systems is to have them not identify as people. But to say that, yes, this engagement was by a generated AI system, because no matter how many times we tell people that they're not real, people consistently say they're conscious. One gentleman in the newspaper said he wanted to marry his AI. You know, like, I mean, there's like this real emotional attachment. And I think that that's very dangerous, especially for children in schools. So that's what this new paper is about. And I'm getting a lot of feedback from people that have read it.

Good. It's really good. I, do you know Zak Stein's work? Have you come across Zak Stein? No, I'm not sure. He's, I've had him on the podcast before. He's an educational philosopher, but I'm happy to connect you with him. But he's recently started an institute for the psychological harms of AI, but he's talking about attachment hacking, exactly what you are talking, I think perhaps what you're calling an embodied hijack, he's calling attachment hacking. And that it's like the kind of cognitive or attentional hacking that, that social media and the, the limited technology was doing. is bad enough, but this whole next layer of this attachment hacking, as exactly as you say, it's like tricking this ancient part of you.

Yes.

That wants to engage with an organism is, it's really dangerous. So no, it's really a good paper. And the thing that I was thinking about when I framed the question was the incentives for the companies are to create more engagement and therefore they're gonna go more down the anthropomorphism kind of design line. But from an educational perspective, if we know it's harming young people, but we also know that AI, these AI companies and systems and platforms are coming into education, it's really hard to know what the responsibility of the educators and the district leaders and the ministers is in relation to that, right? Because it's almost like the two are pulling in different directions.

I think you're onto something because I think schools are just so afraid of using AI-generated text. That's all they think about. Like, you know, my nephew says, well, they use ChatGPT Zero or something that, you know, scans the papers for, you know, AI generation. That's like not even an important piece for me. The important piece is this sort of duping of the public in terms of not even the public, it's our, you know, our brains. It's like our emotions and everything are being hijacked.

Yeah, no, absolutely.

I think we see this too in therapy as well with this reliance on people now to use AI as their therapist. And there are some real serious implications and laws that need to follow in terms of warning people because they do so easily forget that they're talking with an agent and not with a human. And what does that mean for mental health and prescribe— not necessarily prescribing, but just even giving medical information or psychological treatment? And so I think that that is a really— it's a separate issue. But I think here again, where we see the legislation hasn't caught up with the technology.

Yeah. Interesting. It's a fascinating and deeply confusing time, isn't it? Because it's like, it's almost like they're pulling in a different way, pulling in different directions. Because this, what we've just been talking about, about Michael Levin's research and then this kind of allowing cognition everywhere and agency everywhere in even, and then at the same time saying, yeah, but not there, but not in the LLM, right?

Oh, really good point.

Very good. It's really hard. It's really confusing. So I can't even imagine how, you know, teachers are feeling about it, right? So it's, it's It's a challenging time for sure, but we're all trying to like figure it out together, I think. But maybe just, I mean, I'm conscious this has been way too short and I would love to talk to you for a lot longer, but just to kind of bring it to a close, where do you see, like what kinds of hopes, I suppose, maybe do you have for these 2 books in terms of how they provide a bit of additional support or evidence for important changes that, as you've said, Sheila, need to happen, deeply need to happen? in the education systems everywhere. What are your hopes for the work?

I think for me, it's just the lines of communication need to be open. We need to get on the same page with what we're, with what we're discussing. And I think, you know, when we talk about concepts and constructs like intelligence and embodiment, as we've discussed today, we have such a rich history from different fields coming together, but we need first to get on the same page. And so I think both of our books, hopefully that was a goal and hopefully have achieved that goal to some extent, but then just having crosstalk. With our different disciplines, because so much of psychology, of course, was informed by philosophy and computer science is informed by psychology. And we have all of this rich relationship. But, you know, as we know, our academic systems are still very siloed and there just doesn't— there isn't the crosstalk that we need. So I would say a very simple goal of both of these books is just to get fields to communicate with each other. Let's define what we're talking about, which is harder than it sounds. And then let's, you know, let's work together to try to advance the science and the policy as well.

Amazing. Sheila, any final words?

Well, I just want to say how excited I am about the future. And, you know, at this point in my career, I wish it had happened earlier, but I think I'm along for the ride and I'm evolving as a cognitive psychologist. I'm getting into all kinds of things like artificial general intelligence, and I'm talking with people across disciplines, and it's very, very exciting. I mean, I think I think there, like you said, the zeitgeist is out there that there's like sort of this electricity in the air coming together, this collective kind of thinking and moving. And I don't know what's going to happen in the next 5 years, but it's like really exciting if all these different disciplines can, you know, crosstalk and share. I think we can inform each other. And of course, then we also have to worry about AI and that sort of thing. So Like Jennifer said, the policies and things have to be put in place, but I'm very excited about the future. And, you know, this sort of evolution of this cross-disciplined approach to, you know, this new field. I don't even know if it has a name any yet. You know, I don't know what to call it. It's just insane that it's just, it's very exciting time. But I think that we can couch it in embodiment and just move from there. And as soft robotics and hard robotics and computer science and LLMs are starting to adopt this embodiment notion that if we can sort of share the actual understanding of the concepts, that they're going to use them correctly and be able to defend what they're saying, that, that's really important for us. I think that was really a point of the book, that we would have sort of this epistemic alignment across the disciplines, that we could talk about the same stuff. And, you know, that, that's really important.

Amazing.

And I think it's really important too that we have the computational neuroscience too to support these ideas, you know. I mean, all the way from Carl Friston and his students' work that we didn't have, you know, things that we talked about philosophically 20 years ago, even 40 or 50 years ago, we now have the mathematics and we have the understanding to model it and to make sense of this. So I think things that seemed very maybe ephemeral 100 years ago, I think it's really exciting now that we have the science to back it up and we have the technology. And again, that's just a really a testament to the collaborations that we see from computer science all the way through the different disciplines working together.

Yeah. And I mean, it's a critical piece if we're going to persuade the civil servants and policymakers to actually enact some of this into policy. It has to have that level of, I mean, maybe sadly, but it does have to have that depth of deep evidence and maybe decades of development before it really can get trusted somehow.

It's tricky.

It's a tricky one, but no, I mean, I'm glad you're excited. I'm definitely excited too. There's a lot here. I'm so glad to connect with you and just kind of, hopefully there's a community of people out there listening to this who are as excited about this and actually trying some of this stuff in their classrooms and in their schools. And no, I just so appreciate everything you're doing. Thank you very much.

Thank you.

Thank you for having us. Thank you.

We hope you've enjoyed this episode. Please feel free to continue the dialogues with our guests, with us at Good Impact Labs, on our blog or on social media, or within your own networks.