MIT AGI: How the Brain Creates Emotions

In this AI video ...

Today, we’re going to try something different. We’re going to try a conversation. We have Lisa Feldman Barrett with us. She is a University Distinguished Professor of Psychology at North Eastern University, Director of the interdisciplinary Effective Science Laboratory, author of the new, amazing book, How Emotions Are Made, The Secret Life of the Brain. She studies emotion, human emotion, from social, psychological, cognitive science and neuroscience perspectives. And so, I think our conversation may help us gain intuition about how we instill emotional life into future artificial intelligence systems. As Josh Tennebaum gave you a shout out on Tuesday and said that if you want to understand how to create artificial general intelligence systems from an engineering perspective, we should study the human brain. So, with that, let’s have some fun. Okay. So, let me start with the curveball. To conjure up an image of emotion. Have you ever cried while watching a movie that you remember? And what movie was it? I’ve cried during lots of movies. Okay. Let me think. The last time I cried, actually, here’s an interesting thing. When I’m speaking about, when I’m giving an academic talk, I sometimes will talk about a study that we did where we had people watch films. And we have them watch the most evocative clips of films. And there’s several clips that are really powerful. And a couple of them, every time I talk about them, I’m going to try not to cry now. Every time I talk about them, I’m describing for the audience what subjects are seeing. One of the clips is from a movie called Sophie’s Choice. Does anyone know this film? Raise your hand if you know this film. So, this is a film about a woman who is forced in a concentration camp to choose which of her children will die in the gas chambers. And so, I’m already, you know, if I had a heart rate monitor and respiration monitor, you would see, it’s a really powerful scene. Marl Street is Sophie, and it’s very, very evocative. We also show there’s a scene from a film with Susan Sarandon, who is dying of breast cancer. And she has to tell her 12-year-old daughter that she’s dying. And another scene also that I find very compelling. And you use these scenes to listen emotion as part of experiments? We do, yeah. And in fact, I was just giving a presentation to the Supreme Judicial Court of Mestusits on implicit bias. And to start, you know, they want to understand the neuroscience of implicit bias or not they should be crafting jury instructions for juries to be aware of implicit bias. And to open the discussion, I showed them a clip from a film that was filmed, you know, almost 20 years ago, actually, called A Time to Kill, a scene where Matthew McConaughey is this kind of, it’s his closing statements in a case where he is defending an African-American man who murdered two European-American men who raped his 12-year-old daughter. And so the scene is, he’s this completely pathetic lawyer until the end. When he masters this fantastic defense, basically, of this man on trial for, you know, basically avenging his daughter’s death. And that one, I hadn’t seen that for 20 years that film and it brought me to tears, actually. It’s just a really powerful, powerful scene. It just really punches you in the stomach. You just can’t help but just experience the weight of racism in that room and the brilliance of the closing argument to sort of puncture through it, basically. Right. Okay, so experience. So one of the things you talk a lot in your work in your book is the difference in the experience of emotion, the expression of emotion. So what are, what’s our biggest misconception about emotion for folks that haven’t considered, have only considered emotions at a surface level? I would say one common misconception is that you can look at someone’s face and read their emotion the way you read words on a page. That everyone around the world, when they’re feeling angry, they scowl, when they’re feeling happy, they smile, when they’re feeling sad that they frown, the way that I’m saying it sounds preposterous, and it is preposterous. But unfortunately, that is actually what people, a lot of people believe, and tech companies around the world spend tremendous amounts of money. And, you know, in the investment of some of the most creative people on this planet trying to build emotion detection systems, when really what they’re building are excellent systems for reading facial movements, which have no intrinsic emotional meaning. I mean, sometimes facial movements communicate emotion, but many times they don’t. So people scowl for many reasons. They scowl, when they’re concentrating, they scowl sometimes, when they’re sad, they even can scowl when they’re happy, people smile when they’re angry, they smile when they’re sad, they don’t just smile when they’re happy. They sometimes smile to indicate, to send a social message that has nothing to do with emotion. And so, the idea that there’s one universal facial expression, for example, for each emotion is just, there’s no strong scientific evidence for that claim. So how is emotion created then? Because we nevertheless feel like we’re observing emotion when we’re communicating with others. So how does emotion, the creation of emotion change in the presence of others? Does the audience change the display of emotion? Is that essentially in the message? Well, emotions are not displayed, I would say. So basically, you and I are having conversation right now. And part of what, you know, my brain is doing is it’s guessing, making educated guesses about what your facial movements means. So right now, you have a slight smile on your face, not exactly a smirk, but. And so, you know, so, I’m Russian, by the way, so we’re not allowed to show any emotion. I’m trying my best. You know, I have a friend of mine who is Russian, who told me when she moved to this country, actually I’ve had several friends tell me this, that their cheeks hurt for a year, for how much smiling they had to do. I’ve also a friend from the Netherlands who moved here who told me the same thing, her face ached for how much smiling she did. In Embellgaria, they haven’t apparently like a name for, you know, the pervasive American smile, kind of like a not flattering name for how much American smile. But basically, when you look at someone’s face, you’re making a guess about how they feel. And although you yourself are focused on their face, to you, it feels as if that’s where all the information is. Your brain is actually taking in the entire sensory array. So it’s taking in also, you know, the sound of the person’s voice and the person’s body posture and the dynamic temporal changes in all of those signals to make sense of things. It’s not, you’re not really, the face is not a lighthouse, it’s not a beacon that just displays somebody’s internal state in an obligatory way. Nevertheless, there’s some signal there. Of course. And so what are, what would you say, as far as I understand, there’s no good answer yet from science is what are the basic building blocks of emotion? Well, I wouldn’t say that there’s no good answer from science. I would say scientists disagree. I think it’s very clear what the building blocks are. But, you know, in one way or another, if you look back all the way to ancient Greece, you can see that there are two kinds of views of emotion that have been battling it out for millennia. One is the idea that you have, I will give you the modern versions of these views, but they really do have a very long history. One is the idea that you are born with circuits in your brain that are prewired for emotion. So that everybody around the world is born with an anger circuit, a fear circuit, a sadness circuit, a happiness circuit, some other circuits too. You have them, we all have them, all neurotypical brains have them, and actually other animals have them too. And the idea is that when one of these circuits is triggered, you have an emotional response, which is obligatory. So you have a very stereotypic change in your body, your breathing changes in a stereotypic way, your heart rate changes, your chemicals in your body change, in a particular way. You make a particular facial expression, and you have a propensity to make a particular action, like you attack an anger and you run in fear, or you freeze in fear. And then there is another view, which says, well, there are some basic kind of ingredients or building blocks that the human mind or the human brain now. People talk about the brain now, there are some basic ingredients that your brain uses, and it uses them to make every kind of mental event that you experience, and every action that you take. The recipes are what are unique, the ingredients are common. Just like you can take flour and water and salt, and you can make a whole bunch of different recipes with them, some of which aren’t even food, like glue. In a similar way, the idea is that your brain has some kind of all-purpose capacities, and it puts these ingredients together, and it makes a motion as you need them on the spot. And you don’t have one anger, you have a whole repertoire of anger. You don’t have one sadness, you have a whole repertoire of sadness. And if you grow up in a culture that has no concept for sadness, you don’t experience sadness and you don’t perceive sadness because your brain becomes wired to make whatever mental events that exist in your particular culture. So for artificial intelligence systems, the idea of emotional intelligence is really difficult, and it seems like it’s an important thing to try to instill. So for human beings, where do you see the importance on the priority list of what makes us human? Where does emotion sit? I actually think that’s the wrong question to ask, because I think people constantly ask that question, but I don’t think it’s the right question to ask, because some cultures on our planet don’t even have a concept for emotion. And people don’t, while they experience what I’m going to refer to as in scientific terms we call affect, which is our simple feelings, a feeling pleasant or unpleasant, or feeling worked up or calm or comfortable or just, you know, having discomfort. These are feelings that come directly from the internal state of the physical systems in your body. Not everybody makes emotions out of those feelings. And in fact, we often don’t make emotions out of those feelings. The way to think about it is that your brain comes wired to regulate your body. Actually, sorry, take that back. Your brain comes wired. When you’re born, your brain comes wired with the potential to regulate your body. Infants actually don’t regulate their own nervous systems very well. They can’t put themselves to sleep. They can’t calm themselves down. They can’t regulate their own temperature. That’s what they need caregivers for. And as caregivers regulate the nervous systems of their infants, that wires the infant’s brain. So infant, little infant brains are born, like little miniature adult brains. They’re born, waiting for a set of wiring instructions from the world. And they wire themselves to the physical and social realities that they grow in. And as they learn to do this, they experience these simple feelings, that, you know, a feeling pleasant or unpleasant feeling worked out or feeling calm. These are kind of like barometer readings in a way. They’re very simple. And they lack a lot of detail because of how we’re wired. And we have to make sense of them. And the way that a culture makes sense of them is not always about emotion. So I would say, you know, in our culture, when we lose something significant, like a loved one, we feel sad. And to heat, in dehesion, people feel sick. They have an illness. It’s not sadness. A hundred or 200 years ago, there was an emotion called nostalgia, which killed people. For example, after serving in World War One, we would now call that depression. It’s not just a matter of changing the label of something. It’s actually the formation of the experience is very different. So if you want to build an intelligent agent, I don’t think emotion is what you need to endow it with. You need to endow it with these basic ingredients that it can use to make whatever experiences or guide whatever actions, make whatever states or guide whatever actions are relevant to the situations that it’s in. That will be different if it’s an American or a British or a Western, urbanized environment, then say if you were to go to Tanzania and study, you know, the Hadsah who are wondering and gathering since that culture, since the place to see. So do you have words or human interpretable labels on these basic ingredients that we can understand? I think so. I mean, I think the first thing to understand is that when we think about building an intelligent agent, we think about endowing the agent with cognition or with emotion or with the ability to perceive things in the world. Because as humans, these are the things that are really important to us especially in our culture, thinking, feeling, seeing, and other cultures, you know, they have different ways of parsing things. But the truth is that your brain did not evolve so that you could think and feel and see. It evolved. The brains evolved to control bodies. The brains evolved so that they could control the various systems of the body as creatures move around in order to gain resources like food and water. And a brain has to figure out how much resources to expend on getting additional resources. Now that may sound really trivial, but it turns out it’s actually really hard. And scientists actually haven’t really figured out completely how brains do this. I mean, what’s really interesting to me is that if you look, for example, at computer vision, try to figure out, you know, how to make an agent see, that’s a pretty much solved problem. Not completely. It’s a pretty much solved problem in the basics or so. It’s pure perception problem. Yeah, the pure perception problem. However, if you want to create an agent that just reaches out smoothly, grabs a glass and brings it into the body to drink, that problem is not solved. Something as simple as movement, which we think of as this really basic thing like, oh, it’s so trivial, all animals can do it is actually the heart of one of the hardest problems to solve. And brains basically, I mean, there’s a whole story here about evolution, which has nothing to do with having a lizard brain, which is wrapped in a cognitive brain or anything like that. That’s a reference to the trion brain, which a lot of people believe it’s a way of thinking about brain evolution. Instead, what seems to be the case is that as bodies got larger and there were more systems, because the niche of the animal, the environment of the animal got bigger and bigger and bigger, brains also had to get bigger. But they had to get bigger to a point with some constraints. Like, they have to be, you have to keep metabolic costs down. It’s really important. If you don’t, creatures get sick and die. We can talk about what that means in terms of depression or metabolic illnesses or what have you. But so what are the basic ingredients? Well, one of the basic ingredients is that your brain is controlling your body all the time. Whether you feel it or not, whether you’re thinking about it or not, whether you’re asleep or awake, certain parts of your brain are always very active all the time, even when you’re sleeping or else you’d be dead. And those parts of the brain that are controlling your heart and your lungs and your immune system and all of those regions that are controlling the systems of your body are also helping your brain to represent the sensory consequences of those changes in your body, which you don’t feel directly. You don’t feel your heart beating most of the time. You don’t feel your lungs expanding most of the time. And there’s a really good reason why we are all wired not to feel those things. And that is you’d never pay attention to anything outside in the world ever again if you could feel every little movement that was going on inside your body. So your brain represents those as a summary, these kind of simple summary feelings. You feel good, you feel bad, you feel great, you feel like shit, you feel really jittery, you feel really calm. And these feelings are not emotions. They sometimes can your brain can make them into emotions, but they’re with you every waking moment of your life. They are properties of consciousness in the way that lightness and darkness is a property of vision. And sometimes we make them into emotions when they’re when we have a big change in our heart rate or a big change in our in our breathing rate or a big change in our temperature or a big surge of glucose. We might the brain might make emotion out of those changes, those very strong changes which you will feel as really feeling unpleasant or really feeling pleasant. But your brain might also make hunger or your brain might make an instance of a physical sensation like nausea or your brain might even make a perception of the world. Like, that’s a nice guy. That guy is an asshole. This is a really delicious drink. That’s a beautiful painting. Those are also moments where these simple feelings which we call affect are very strong. So affect is a basic ingredient. There are others that I could talk about too. It’s not the only one. But one of the things I think that sometimes people who are studying to build AI systems don’t realize is that the brain, it’s fundamental job is to keep your body alive and well. And if you don’t have some kind of body to regulate with affective feelings that come from that regulation or something like that, it’s you’re kind of going to be out of luck, I think. In rendering something that looks more something that seems more human. Right. So maybe you can elaborate like in the book, Sapiens that as human beings are really good on mass as a thousand millions of people together believing something even if it’s not true. So, while scientifically sort of from a neuroscience perspective, it’s maybe very true that emotions aren’t real. I didn’t think they aren’t real. But I said they’re not there. There isn’t. So it’s an interesting, you finish. So what I’m trying to say is also from AI perspective is they become these trivial ideas of mapping a smile to being happy and these kind of trivial ideas become real to us through Hollywood, through cultural spreading information and we start to believe this and therefore becomes real. In as much as anything is real about our perception together. So it’s really important scientifically the ideas that you’re presenting. But does that mean there’s just because our brain doesn’t feel those explicit emotions. Does that mean they’re not real. I didn’t say we don’t feel explicit emotions. So I want to be really clear about this because it’s an interesting this, you know, this. This inference, it’s an interesting inference that people often make and so, but it’s it’s a mistake and it’s a mistake that betrays a certain kind of thinking that we do in this culture. And it’s it’s the mistake of the following sort. So when I say there’s no facial expression that is diagnostic of a single emotion. That doesn’t mean that people don’t express emotion. They certainly do express emotion. They just don’t, you know, when when they’re in a state of anger in a state of sadness or in a state of awe their faces don’t do one thing. When I say well your body can do many things when you’re angry or when you’re sad or but let’s take anger your heart rate can go up it can go down it can stay the same your breathing rate can go up it can go down it can stay the same. The same pattern that you see in anger you sometimes see in sadness and you sometimes see in fear you sometimes even see it in enthusiasm and in awe. So does that mean that emotions aren’t real no emotions are real but but sometimes things are real because the physical meaning of the signal is endowed in the signal. Okay so when your retina communicates to your brain that you have that you are faced with a wavelength of that you know is 600 you know 600 nanometers. The signal is endowed in your brain that’s like in the signal it’s not like interpret interpret that it’s 600 you know nanometers it is 600 nanometers the information’s in the signal. But when you see red that information is not in the signal. Your brain has added information that isn’t in the signal itself in a sense your brain has imposed meaning on a signal that the signal doesn’t have on its own. So let me back up and give a different example to make it a little easier and then we’ll re-approach this. There’s a there are some things that we impose in fact this is true almost of all civilization right that we there are some things that are real by virtue of the fact that we agree that they exist. Little pieces of paper serve as money or now you know Bitcoin or little pieces of plastic or gold or diamonds or in the past barley salt shells rocks serve as currency have value only because a group of people agree that they have value. So we impose meaning on objects and once we all agree that that object actually has value we can trade it for material goods the minute that some of us disagree that we were we withdraw our consent right. The things lose their value that’s what happened in the mortgage crisis that’s what happened in the tulip crisis in the in you know the Netherlands in the 17th century or 16th century or when it was. Money currency exists because we impose meaning on objects in the world physical objects in the world that themselves don’t have that meaning on their own and they are very real money is very real to people I can stick somebody’s head in a brain scanner and show you that they experience value in a very real way but that that reality is constructed. By the fact that they have learned the value in a particular culture well that’s also what we do with emotion we impose meaning on certain physical signals that they don’t have. By on their own and but we have collective intentionality we all agree that scowling is sometimes anger and so it becomes anger in a very real way just like little pieces of paper become money so it sounds like you kind of think about the expression of emotions of kind of language is an extension of language that we learn in the same way that we collectively agree on a language on the lexicon and how we use that language. Sure you could think of it that way I mean everything everything in our culture. Is almost everything in our culture is a function of social reality in this way we are citizens of a country because we all agree that the country exists more or less and wow you guys barely even laughed at that okay. What’s a revolution a revolution is when some people in the country withdraw their consent they no longer agree. Right a president has powers in a country because we all agree that a president has powers present only as powers by virtue fact that we all agree that he or she has powers if we stop agreeing. The president doesn’t have those powers anymore it’s very real people’s lives depend outcomes of real people depend on these social realities that we build and nurture and we why are these social realities into the brains of our children as we socialize them and when people move from one culture to another they have to learn the new. Social reality that they are faced with and if they don’t they get very sick physically because our ability to agree on what something means actually is important for regulating our nervous systems so can you speak to that a little bit for machine learning methods for. System that learn to behave based on a certain reward it’s important to kind of have some ground truth and learn so how do it sounds like the expression of emotion is learned. Can you talk about how we learn to fit into our culture by expressing emotion with our face body given the context given the rich what’s that process look like when does it happen. Sure well you know I mean I wrote a 400 page book so I’ll try to do it in like a couple sentences yeah. So so how does it how does an infant learn anything so an infant’s born. And it can’t really do anything for itself. It can’t regulate its own nervous system it can’t keep its body systems balanced this is a term scientific term for this is alastasis alastasis is your brain’s ability to predict what your body’s going to need. Before it needs it and tries to meet those needs before they arise so an example would be if you’re going to stand up if your brains going to stand you up it has to raise your blood pressure before it stands you up. If it doesn’t you’ll fall that’s costly from metabolic standpoint it’s costly you’ll hurt yourself. So an infant’s brain can’t do this very well. An infant doesn’t know when to go to sleep and when to wake up. An infant doesn’t can’t feed itself can’t regulate its own temperature. Someone else has to do it. And when someone does it the infant is learning the infant is learning. It’s taking insights and sounds and smells and the physical sensations from the body which are comfortable and pleasant. When the infant alastasis is maintained. So right from the get go an infant is learning statistical learning. You know the capturing events including their consequence for the infant’s body. Some people think babies are born you know attached already to their caregivers but they’re not actually. Infants don’t even know what a caregiver is it’s just that the caregiver is there constantly meeting that infant’s needs. That’s how infants start to learn. Now if there are statistical regularities like for example an infant is not born with the ability to recognize a face as a face but it learns that in like the first couple of days of life. Why? Because human faces have some statistical regularities to them right two eyes and nose and mouth kind of in the same place most of the time. So it learns really quickly. But here’s the interesting thing around three months of age. Infants start to learn what we call abstract categories. They start to learn that some things which don’t look the same or sound the same or smell the same. Actually have the same function. And how do they learn this? They learn it with words. So if you do an experiment with a three month old three months old okay and you say to that baby very very intentionally look sweetie this is a wug. And you put the wug down and it makes a noise like a beep. And then you say I’m like I don’t have props. And then you say I’m gonna wall it. And then you say give me your wallet. And then you say look sweetie this is a wug and you put the wug down and it makes a beep. If you say look sweetie this is a wug that infant expects that object to beep. Why is that important? Because in the real experiments this might be yellow and squishy and tall. And this might be red and pointy and hard. And this might be you know so lots of different perceptual features. But the infant knows learns that the word is inviting the infant to understand that the function of those very different physical signals are actually the same. Similarly you can take you know six objects that are exactly identical in their physical features. They sound how they smell what they feel like what they look like and you can name three of them with one word and three of them with the other word. And the infant will understand that these are actually different objects. That happens a little after you know not as early as three months. But the point is that we talk all the time. We use words all the time. What do we do with infants? We’re constantly pointing things out and labeling them. This is a dog, this is a cat. You’re angry. This person’s sad. Oh mummy’s really happy today. Oh you know mummy loves you. Oh daddy’s really excited about this and so on and so forth. And infants learn really really quickly. Words are considered to be kind of invitations to form abstract concepts. That is the basis of almost all of the you know mental categories that we that we mental events that we experience. We’re basically teaching children to form these abstract categories. And that’s the basis of money and it’s the basis of rules that we have with each other and what we expect from each other. It’s the basis of a lot of the sort of functional categories that we use in everyday life. We’re imposed meaning on sensory on sensory arrays that those sensory arrays in and of themselves don’t necessarily have. They only have that meaning because you and I both learn that that package of sensory array means something. So when I make that you can anticipate what will happen next just because we’ve learned those are either wired into our brains you know in our culture. And when we go to a different culture we have to learn sometimes different packages. Yeah different mapping so you’re saying that there is a few sources of sensory data and a few building blocks inside us the feelings of some kind that we learn to then from an early that we’re come born with those. Part of what your brain is doing is it’s trying to make sense of the sensory array around it. So from your brain perspective, from your brain’s perspective, just think about it from your brain’s perspective. Your brain’s perspective, it spends its entire life trapped in a dark silent box. And it has to make sense of what’s going on around in the world so that it knows what to do to keep itself alive right and well. But it has to know what to do based on it has to know what what’s happening on around it only from the effects that it receives through the sensory systems of the body. So a flash of light, what’s a flash of light it could be anything. What’s a like a siren, a siren could be a fire track or it could be somebody’s car alarm went off or it could be a doorbell or it could be right any particular sensory cue could have multiple causes. So your brains trapped basically in your skull and all it gets are the effects, the sensory effects of stuff that happens in the world, but it has to figure out what those things are so that it knows what to do. So how does it do that? Well, it has something else that it can draw on. It has your past experiences. Your brain basically is doesn’t store experiences from the past. It can reconstitute them in its wiring. And that’s what it uses to guess at what those sensory cues mean of those sensory changes mean. So in one situation, a siren means one thing in another situation it means another flash of light means one thing in one situation and a different thing in another. So your brain is using past experience to make guesses about what these sensory changes mean so that it knows what to do. And it has the same relationship to the sensory changes in your body. What’s an ache in your stomach? Well, it could be hunger. It could be anger. It could be disgust. It could be longing. It could be nausea. It’s not that there’s one ache in your stomach for nausea and another ache in your stomach for hunger. There are many aches in your stomach, many different feelings of achiness for nausea and many different feelings of achiness for hunger and sometimes they overlap. So your brain has to make the same kinds of guesses about what’s going on in your body as it does about what the sensory events mean in the world. And that’s really what it’s doing. It’s guessing and making sense of the sensory array so that it knows what to do next. And when it guesses wrong, it takes in that the sort of information that it didn’t predict well, which in psychology we have a really fancy name for that. We call it learning. Your brain takes in the information that it didn’t predict so that it can predict better than the next time to make sense of things the next time. So you kind of answered this a little bit. I’d like to elaborate on it. If you were to build a robot that performs maybe passes the joint test or performs at the low bar level of instead of myself here today, it would be a robot talking to you. It would be convincing as a human. How would you, what kind of, how would you build that system in a sense in paralleling the infants? What essential aspect of the infant experience do you think is important? It needs to well, I mean, you know, I don’t, I’m not a computer scientist. But so the way that I would say it is it needs to have a body needs to have something like physical systems or an analogy to physical systems. It has to do something analogous to Alice Stasis. So what’s the site to elaborate? So what would be the goal for the system? You kind of mentioned previously that the goal would be for the brain to just stabilize itself. So people talk about reward. What is reward? The machine learning is pretty easy. It’s something. It’s mathematical. So there’s no philosophy to it. Oh, there’s philosophy to everything. Right. Whether you admit it or not is a different story. Yeah. So you wanted to play game chess, play game, go. You wanted to pick up a water bottle is a, okay, but what is reward? Existence? Don’t mean it’s actually not reward. Don’t mean it’s effort. Don’t mean it’s necessary for effort. It’s not necessary for reward. I mean, it’s commonly, if you read the most, most up-to-date literature, that is what you’ll see. That is actually, people can, animals can find things rewarding, can be without, without dopamine, actually. They need dopamine to move. They need dopamine to encode to learn information. So it’s really for effort that is required to work towards getting a reward, I would say. When a brain, an animal brain, any kind of animal brain, mispredicts what the reward will be. That’s what you see a real surge of dopamine, because the animal has to adjust its action. But reward is basically bringing the body back into alistaces. It feels good when that happens, and people will, and animals will work tremendously hard to have that happen. So, what is motivation? Motivation is expanding resources to get a reward. So basically, if you don’t have something like physical systems that have to be kept in balance, water, I mean, for humans, or for actually any living creature on this planet, even single cell organisms, actually, there’s an analogy to what we’re talking about here to brains. But salt, water, glucose, all these systems have to be kept in balance, and they have to be kept in balance in a very, very efficient way. And that’s the motivating, so-called motivating force, really. That’s what really brains are for. And that is the consequence of that regulation, it are the basis of affective feelings, which are for many, many creatures on this planet, a property of consciousness. Okay, so maybe if it’s okay, we’ll take some questions from the audience. But first, let me ask the last question, so I’m building on the robot question, how would you build the same kind of robot that you would be able to, as a human being, fall in love with? Well, people fall in love with their cars. They fall in love with their blankets, they fall in love with their toys. You know, it doesn’t need much to fall in love with something. The question is, will it love you back? No, well. I would elaborate, I think, I think you’re answering that love in the way we’re defining it loosely in poetry and culture as a social construct, and it’s relative. What I mean is sort of the idea of monogamous long term love that we have deep connection with other human beings that we have. You’re saying you could do the same with a car, like a nice 69 most things. Are you telling me that you’re telling me that you’ve never, you don’t know anyone who was like so in love with their car, that you, okay. Now, here’s the thing. So one, so here’s the thing. We are social animals, okay. What does that mean? What does it mean to be a social animal? It means that we regulate each other’s nervous system. So our brain, my brain isn’t just regulating my nervous system right now, it’s regulating yours. And actually, it’s regulating other people’s in the audience too, and vice versa. And why is that? I mean, other animals do it too. We’re just really good at it. But other animals, like there are some insects that are social species, they regulate each other’s nervous systems, they do it with chemicals, they do it with smell primarily, and a little bit with touch. Like earwigs, well, you know, like they actually, I have this great picture of this like totally disgusting looking little bug, but it’s like, you know, cuddling its little baby, ugly, ugly baby bug too. It’s a adorable picture. But, you know, what about mammals like rats? Well, they also use chemicals like smell, but they also use touch. And to some extent, they also use sound, they use hearing, primates ad vision. And as primates, we do all, we use all of those senses to regulate each other. And we also use something else. Words, right? Exactly. And words, the systems in our brains that allow us to speak and allow us to understand words are directly connected to the parts of the brainstem that control the body. I don’t mean like there are a bunch of, I mean monosynaptically connected. So exactly the same systems in your brain that are important for you to be able to understand language and to speak are also directly, directly affecting the systems of your body. And that is why I can say something to somebody, I can speak to you and I can have an impact on the nervous systems of people all the way at the back of this auditorium without they, you know, maybe they can see me, maybe they can’t, maybe they can’t, hopefully they can’t smell me, maybe they can hear me, maybe they can, you know, but they can, if they hear me speak words, I can affect their nervous systems. That’s why a telephone works, that’s why a telephone works where you can feel connected to someone just merely by hearing their voice, because the sound of their voice has an effect on your nervous system, it can, it can make you breathe faster, it can make you breathe slower. And the words also have an effect, because when I say a word like, I don’t know, when I say word like car, that’s a short form. I have a bunch of mental features in my mind when I say the word car, and I say that word, and that invokes those similar mental features, maybe not identical, but similar enough, they invokes it in your mind. And your mind is made by your brain, so it invokes, if I just say the word car, there are changes in your motor system that would be exactly the same or very close as if you were actually in a car, right? So this is something that we do, and the fact where attachment comes from an infant to a caregiver or two lovers or two really close friends comes from the ability that we have to regulate each other’s nervous systems. And that is why when you don’t have that kind of attachment, you die sooner, on average, seven years sooner, loneliness kills. I always tell my daughter, my daughter’s 19 years old, and I always tell her, you know, breaking up, when you break up with someone, it feels like it will kill you, but it won’t. Loneliness, however, will kill you. It will kill you. On average, seven years earlier, then it would, if you didn’t have an attachment, and that’s because our nervous systems, you know, as our bodies got really complex through evolution, and our brains got bigger, they could only get so big. There are constraints on how big any brain can get that have to do with, you know, birthing the infant, but it also has to do with the metabolic cost of a brain. Your brain is really expensive, my brain really expensive, three pounds, 20% of your metabolic budget, that’s a lot. And so what did evolution do to solve this problem? Well, it couldn’t make our brains any bigger, so it just entrained other brains to help manage our nervous systems. So you bear the burden of other people’s alastasis, and they bear the burden of yours, not always at the same time, but that’s what it means to give people support. When someone, when you’re feeling horrible and somebody paths you on the back or says nice words to you or gives you a hug, they are physically having an effect on your body. That they are helping your body to maintain alastasis at a time when your brain probably couldn’t manage it on its own. And so the basis of love or attachment is basically that. It’s the ability to affect each other’s nervous systems in a positive way. I always say to people, you know, the best thing for a nervous system, a human nervous system is another human. And the worst thing for a human nervous system is another human. Because we’re social animals. Well, beautifully put. So maybe a few questions in the audience. Do you mind? Does microphones on both sides? Go ahead. Hi, thanks for talking here. You’re saying cool stuff. Thanks. It’s not a question. It’s okay. I’ll take it. It’s all right. So I was thinking about what you’re saying about reward. And I’m wondering, first of all, is he described as a return to alastasis, is reward in any way linked to the, I just like the reinforcement of pathways or behavior so that the next time you get that stimulus, you will, you’ll respond in the same way. And I’m also wondering about the link between the desire for alastasis and the need for novelty and the need to explore environment. Yeah, it’s a great question. It’s a really great question. So the second question is so much more interesting than the first actually. So let me say this that there is a need for novelty. The need differs for different people. I will say novelty. So first of all, the, we can think about the need for novelty in a really proximal way, or we can think about it in a really distal way. Like, but basically, when I say that a brain is organized or engineered for metabolic efficiency, that doesn’t mean that the goal is to only ever have your prediction, you know, your brains predicting all the time. To always have its predictions completely perfect so you never learn nothing because you’ll be bored out of your mind, right? And also, you know, humans like to expand their niche, they like to explore. So it’s a constant balance between what biologists would call exploitation and exploration. Novelty is, is it’s exciting. There’s actually an increase in neuropinephrine and increase in arousal. It feels really exciting. But it’s also super costly. Novelty requires usually that you learn something new. That means that’s actually really metabolically expensive thing to do. And it also means usually that you’re moving your body around, which is also metabolically expensive thing to do. So the need for novelty is balanced by its cost. And different nervous systems can bear different amounts of cost. So for example, if you take two rats that are somewhat genetically, you know, like they’ve been genetically bred, one is bred. When you stick it in a novel cage, it just sits still. And the other one, when you put it in a novel cage, it like roams all over the place. And it’s just, you know, it’s going crazy kind of exploring everything. Well, the one that sits still, scientists might say, well, that’s a nervous rat. Or that rat’s afraid. What is that rat doing? The rat is not moving. And it’s not encoding anything because encoding something is expensive. It’s expensive. This rat on the other hand is roaming all over the place, moving a lot, learning a lot. So it’s encoding a lot. Spend, spend, spend. Save, save, save, spend, spend. There are differences between people and there are also differences between times in your life, where moments where you feel like you have a little bit to spend, and other moments where you feel like you really have to conserve. When I talk to the public, I always talk about, I don’t use the word Alice Dases. It’s just too boring a word. But I sort of do is sort of explain it like a budget, you know, like your brain is sort of like the financial office of a company. Has lots of offices. It has to balance the expenditures and revenues. And it’s got to keep everything in balance. So it might take a little money here, move it over to that office. You know, it’s got to keep everything in balance. What it’s always trying to do is spend a little bit to make a little bit more. What happens when it spends a little bit and it doesn’t get a revenue back? There’s no reward. What happens? Well, it goes into the red a little bit. So what do you do when something goes into the red? Well, you might do something risky. You might actually spend a lot to try to really make, you know, not just make back your deficit, but actually make a lot. That would be novelty. That would be move and spend, move and encode. Or you might reduce your spending. You might say, well, I’m going to save a little bit now. That would be, I’m not going to move too much. I’m not going to spend too much. I’m not going to encode anything. So I certainly don’t mean to suggest to you that that novelty is unimportant or that learning is unimportant. And it’s a really important question about what is there any intrinsic value to novelty over and above the rewards that it would give you in an allostatic sense. But it is really clear to me that the extent to which you any nervous system will embrace novelty and even seek it pretty much depends on the allostatic state of that system. If you don’t have a lot to spend and you’re already in the red, you, at a certain point, if you continue to spend when you’re in the red, you go bankrupt. What that means in human terms is you get depressed. It means that your brain makes you fatigued so you can’t move and it makes you locks you in so you stop paying attention to anything going on around you in the world and your experience is just what’s in your head. That’s actually what depression is. So that makes me think of all the stasis more of as a range than as a zero point. It’s not homeostasis. It’s allostasis. It’s a range for sure. We’ll answer some other questions and maybe I’ll get back to your first question if there’s time. So if I understand your argument correctly, if we’re going to make anything like a general intelligence, something approaching, you know, like a human, it needs to be an embodied system. I want to be careful about saying that because for two reasons. One, because in biology, there is this concept of degeneracy, which is a sucky word, but it’s a great concept and it means that there’s more than one way to skin a cat, basically. You want a functional outcome. There are many ways to get to that functional outcome. For example, there are a lot of characteristics that are heritable, but we don’t know the genes for them. And the reason why is that there isn’t one set of genes. There are multiple sets of genes that can give you actually the same outcome. So what I want to say is that you need something akin to a body. It doesn’t actually have to be a body. I imagine there are lots of ways that you could implement a system. You could implement an agent that has multiple systems of some sort that it has to manage. My point is that one thing about that is very important that we continually miss. When we think about building an agent with mental states, we continually miss the fact that it has a body. It has a human’s have bodies, and that’s the brain’s primary task. And our most fundamental feelings come from the physical changes in our body, even though we don’t normally experience it that way. That actually is how it is if you just look at the wiring of the brain, you can just see it. So it seems to me that if you want to build an agent that is human like it has to have something like a body doesn’t have to maybe be a body. And I’m sure there are many clever ways that you could implement something like a body without it actually being a body. If you understand what I mean. So Amazon Alexa could be there if we just gave it some sort of representation of mental states or some kind of aloe static. Yeah target. Sure. Let me just say one other thing. All brain requires is that you at some point had a body. You know, right. So basically, this is what phantom limb pain is. This is what chronic pain is. This is what happened. If at some point you cease to get information from your body anymore, your brain still can simulate still can reinstate the sensory patterns that once came from the body. And that’s what’s required. Right. At some point, the body isn’t really needed anymore. Yeah. I think we’re here. Yeah. So emotions in humans look as if they implemented by a bunch of hacks. So there’s a bunch of chemicals that go into brain. There’s moxytocin serotonin. A whole bunch of biochemical things that diffuse around in the fluids in your brain that affect your emotional state. And that seems like a hack that we’ve inherited over millions of years from primitive ancestors. And if you look at the machine learning world, we can do a bunch of similar things with neural net so you can increase the activation thresholds on a large scale. Yeah, change the map noise going into the system. You can do a bunch of similar things, but you don’t have to rely on fluids being kind of cleaned slowly by glial cells. Things don’t diffuse around in the fluid in the system necessarily. Seems like there’s a lot more flexibility. So when you come to implementation, there’s not so many constraints imposed by the evolutionary history on the whole system. And it seems like that would make it work better. So when people are in negative emotional states, they can’t think straight. Oh, that’s absolutely not true. People can think actually quite well in negative emotional states. I have to tell you. Emotions, they can plan crimes, they can do very nefarious things very, very effectively. Right. So, but the general point I think is true, even if that example is not so emotions kind of flood your nervous system with emotions don’t flood your nervous system. So some of them do. So, yeah. So, let’s talk about let’s talk about this chemical. Let’s talk about this chemical. There is not a single chemical in your brain or anywhere in your nervous system that is for emotion. Serotonin is not for emotion. Don’t mean it’s not for emotion. Oxytocin is not for emotion. Even opioid, you know, even even opioids are not for emotion. So, they influence your emotions. They’re involved in emotion. They influence every mental event that you have, not just emotion. So, your brain, for example, your brain is a physical, it’s a set of physical cells that are bathed in neurochemical systems. And the neurochemical systems that you’re referring to basically change the ease with which information is passed back and forth between those neurons. That’s always true. It’s true regardless of whether the event is emotion or whether it’s a perception or whether it’s a thought or whether it’s a belief. It’s always true. So, for example, serotonin, serotonin is a neurotransmitter that allows your brain to delay gratification of a reward. It allows you to expend energy now because you anticipate a reward at some point in the future. And if you have a deficit in serotonin, then you can’t do that very well. And it turns out for humans, one of our great superpowers is the ability to do mental time travel. To remember in the past and also to do things now because we know they’re going to have an effect 10 or 20 or 50 or 100 years from now. So, when I said you have to have something like a body, I’m not saying literally you have to have a physical corporeal body. You have to have, I mean, it’s just a fact. It’s not an argument. It’s a fact that your brain, that brains evolved for the purposes of regulating multiple systems. And from a cybernetic standpoint, you know, the best way to regulate a system is to build an internal model of it. That’s what your brain is. And then the brain is an internal model of your body in the world. It’s running simulations. It’s running this model. And if you want to have an agent that is somewhat human like that has feelings like humans, then they have to do something similar. Whether it’s a actually physical corporeal body or not, I think is, you know, that’s an open question, right? So, it sounds as if you’re disputing my premises before I got to my question. So, I started off by saying, I started off by saying emotions are implemented as a bunch of hacks. So, would you say that was rubbordly correct or would you say that they’re not hacks, they’re finely tuned and adaptive? Well, I wouldn’t say they’re not maladaptive. But would you say it’s a bunch of hacks? I don’t think they’re, I don’t know what you mean by hacks. I think. So, clu-clu-gees, kind of historical accents of your face. I think it’s the wrong, you’re talking about emotions like they’re, they’re talking, the premise of your question, I can’t answer your question because I think it’s not the correct question. Emotions aren’t like blood pressure, you know, they don’t exist in that sense. They are, they are the way that they are, first of all, not all cultures have, not all people in all cultures have emotions. Everybody has affect, assuming they have a neurotypical brain of sorts, but they don’t all have emotion. And so, to answer your question that you’re asking me is, I can’t answer it because I don’t think it’s the right question. Opioids and the pain systems seem like things that influence your emotions fairly directly. So, it’s not that it’s the same thing, but it’s that there’s a powerful link. Opioids though. So, opioids are important for instances of emotion, but they are also important for every other category of mental event that you brain can make. Well, thanks too, I agree. So, I, you know, you will see sometimes scientists will assign, you know, an emotional meaning to something like dopamine is for, you know, first hedonic pleasantness and then it’s for reward. So, I would like to assign single functions to biological entities like a chemical or a brain region, a cluster of neurons, or it’s just not, that’s really, I’m not saying every chemical does everything, but whatever, whatever opioids do, they do in every waking moment of your life, not just in moments that are emotional for you. Thank you so much. Thank you. Yeah. I think it was very interesting that you brought up the last topic of love because I think it actually brings up a really important thing, which is, and what you were saying about connection. That’s, I mean, the primary purpose of life is to procreate. Right. I mean, that’s what our genes do. So, first of all, right? I’m not going to be touching that. Yeah. So, that, and it’s, of course, in humans, it’s not just genetic, it’s mimetic. I mean, we, we procreate our ideas, as well as, you know, physically. And so, a primary purpose of our wanting, our reactions and interactions with other people and other things is that goal is that we get rewarded when we interact with other things in a way that creates something new, whether it’s art or, you know, a book or technology or something like that. And if we don’t put that, if there’s no inherent reason for a robot or a computer system to want to do that, I mean, how, how can we, can we even imagine putting that into a system inherently that it wants to, that it just desperately wants to make something new. Well, here’s what I would say, you know, when we’re, when, first of all, when we talk about, we’ve been focusing a lot on bodies, I’m certainly not saying that’s a sufficient condition, right? I’m just saying it’s a necessary one or something like a body. But I certainly have the motivation to create, and I’m, I’m imagining that you do too. I have to tell you not everybody has that. I’m just out of lessons don’t have that. Many adolescents don’t have it. Okay, that was a joke. I have a teenage daughter. I’m just telling you. The point is that what people, what people find rewarding is remarkably diverse. But the property, I think, is that there has to be a feature of reward. For, for a lot of people, that’s, you know, that’s having an impact in some way. Having an impact on another person, having an impact, you know, on a building something that wasn’t there before, whatever, you know, in innovating or discovering. But it’s not true for everybody. It’s just true for like a lot of people in this room probably. And certainly the people that we probably spend a lot of time with, but not for everybody. And then there’s also the genetic part, you know, sure. I mean, look, but if you want to, you know, if you want to, if you, you know, we can certainly make a genetic argument here, absolutely. And there, there’s nothing that I’ve said today that is inconsistent with the idea that, that you have to, that you have to pass your genes on to the next generation. And actually make sure that that generation survives to reproductive age. It’s not just enough to have offspring. You have to make sure the offspring survive to reproductive age. There’s a whole argument about, you know, the learning of social reality concepts, the social reality that we’ve been talking about, like, money and emotions and so on. That makes that argument, right, that it’s very expensive to have to encode everything in genes or to have to learn everything from scratch every generation. So instead, what we have is a system, a genetic system that allows us to wire the brains of the young with what we know. And that’s what we do, basically. All right, so I’m curious about your analogy about intentionality that you talked about when you use the analogy between money and the perception of red to the facts that we have, the fact that we have emotion. Because the distinguishing feature it seems to me is the level of intentionality. And as you said before, our brain assigns meaning to things. But we don’t, or maybe, and my question is whether or not you agree with this, I guess. We don’t always deliberately assign meaning to. This is nothing I’ve said is about ever deliberate. Sometimes we do, right off and actually. So when you go back to the question of what makes something intelligent, a lot of previous talks have been about, you know, we want to pick a goal and then we create costs to achieve that goal. But that goal is deliberately assigned. So when you talk about what makes something intelligent, what do you think the role of intentionality is and the spectrum they’re in? So first of all, when you talk about intentionality, you have to really be careful that you are philosophers talk about intentionality in two ways. They talk about intentionality to mean a deliberate action, the way you mean it, but intentionality can also mean that something has a reference out in the real world outside of you. So that word, a word has a reference that’s the intention of the word, basically. So I think you have to be really careful. I also think that you have to make a distinction between a conscious, deliberate, explicitly, a goal that you can explicitly describe. And I mean, you’re sort of making, you’re sort of making, I mean, the kind of question you’re asking is getting very close to the question of free will, which I would love to not have to discuss. But basically, and what I’m about to say is going to sound very Cartesian, unfortunately, because that’s English. I don’t know. There’s no other way to do it, actually. But what I want to say is that your brain is always, there’s always a volition, but it’s not always consciously experienced by you as as agency or will. So you’re not a, you’re not a sea urchin, you know, your sensory neurons are not hardwired to your motor neurons. You have inter neurons. That means you have choice. Do you consciously experience yourself as making choices all the time? No, you don’t, but your brain is actually making choices all the time. That’s why people who study decision making, you know, think they’re studying the whole brain because they are actually. So I think you have to be really careful about. There are words that we use in English and in science that have two meanings. They can have a meaning that is about decision making or choice, that is just obligatory, automatic, and a function of how the system works. And then there’s the kind of choice that feels deliberate and effortful and where we feel like we’re the agents, we experience ourselves as the agents. And intentionality usually can be assigned to the second one, but actually in truth, in philosophical terms, it can also be assigned to the first. That when you, even if you’re completely unaware of having made a choice, you’re acting on something with some degree of volition because you, you, it’s not a reflex. It’s not like, you know, somebody hit your patella or tendon and you hit. And so, you know, I think you just have to, you have to make that distinction and I probably should get to the next question we can follow. Thank you. Yeah. So you’ve been asked a lot of ascetic questions about AI, but I think we might gain some insights by wondering about DI that is a dog intelligence. Sure. So I believe I, I sort of understand what my dog is feeling and I usually believe that my dog, you know, believes the same. But not the same with, you know, a cat. Let’s put you not a cat person. Yeah, sure. And, you know, I’m not, I don’t know that much about monkeys, but I’ve never really seen a monkey be able to make the expressions that a, let’s say a dog can. And I was wondering if you had any insights about, you know, why dogs are able to do this and why we’re able to read dogs. Is it something just simple like they’re, they have the right facial muscles or is it something some drive that allows them to learn this. So I think before I wrote my book, I would have answered this question differently. But now here’s what I would say. I think what, you know, many creatures on this planet have affect. Right. And we can debate about whether, you know, I just came across a paper the other day about whether fly, you know, just just Dressoffila have affect. You know, it’s, I mean, it’s actually a really interesting question. They certainly they have something like opioids and they, you know, so it’s a question. But dogs, dogs are really interesting because they do seem to have some capacities that only, that you only see in in great apes. And they may have capacities that great apes, even other than us don’t have. And I mean, they certainly have some capacities we don’t have either. But here’s my point. We actually bred them. We bred these animals. We selected them basically not natural selection. It’s, it’s, you know, artificial selection. And we selected them for a couple of things, right. If you look at the experiments on breeding, taking a fox, you know, taking foxes and breeding them into what look like little dog like animals, it’s interesting what they can do. And one of the things they can do is they can move their facial muscles in a lot of ways that they have a lot more control over their facial muscles than, you know, chimpanzees and do and other apes. And they also do, they do joint attention really well with gaze. So this is something that really no other animal can do, I think other than a dog, other and humans. And that is they can, they, they meet your gaze and they use gaze for reference. So, you know, they’ll look at something and they’ll look back at you and you, you know, that’s actually partly, that’s how we communicate with each other. Chimpanzees lose that ability after about 10 or 11 months of age. But, but dogs continually do it. And actually joint attention shared gaze is how we communicate with our infants. And so, also, that’s actually partly how we teach infants about what’s important in the world is with gaze. So, I think that dogs may actually have some capacities probably because we bred them to have those, but we selected the animals that, you know, and bred them to have those capacities. That’s very interesting. Thank you. Awesome. Well with that, let’s give Lisa a big hand. Thank you so much.

AI video(s) you might be interested in …