The psychology of morality, perversion, and revenge, in 67 minutes | Paul Bloom: Full Interview

I'm Paul Bloom, a professor of psychology at the University of Toronto and an emeritus professor at Yale University. I study human nature — children and development — with a particular interest in morality, the social emotions, and how we understand ourselves and relate to other people.

When we think about loneliness, we tend to think of being alone, of solitude. Solitude can certainly be a source of loneliness, but many of us are comfortable being alone. You can be alone for a long time without feeling lonely, as long as you know there are people who love you, understand you, and appreciate you, waiting in the wings for when you come back.

Loneliness, Solitude, and Empathic AI

What loneliness is at its root is a lack of human connectedness — the feeling that you aren't loved, you aren't needed, you aren't respected. This could mean you have nobody in your life, but it could also mean you're surrounded by people and not connecting to them. Hannah Arendt pointed out that the most painful form of loneliness is when you are surrounded by people and feel utterly isolated. I'm interested in loneliness in part because it's an exquisitely painful experience — one of the worst states a human can be in, to feel alone and disconnected.

A solitude epidemic, not a loneliness epidemic

People talk about a loneliness epidemic. Derek Thompson has a wonderful discussion arguing there is no loneliness epidemic. In fact, the data show people were much lonelier a few years ago during COVID than they are now, and there's not much evidence people are more lonely now than they were 20, 30, 40, or 50 years ago. What we see instead is a solitude epidemic: people are spending much more time alone than ever before.

This makes sense when you consider how technology has improved. Our houses are better, our TVs are better. The average person used to go to the movies multiple times a month; now we watch movies at home on big screens with an abundance of streaming channels. We used to go out to eat a lot; now much of our dining is alone, with orders from Uber Eats. As the possibilities for a rich life indoors, by ourselves, have gotten better, we have become more alone. Whether solitude in itself is a problem is debatable. But regardless of whether loneliness is an epidemic, and regardless of how bad it is relative to the past, it's a terrible thing and many people suffer from it. Estimates are hard to pin down because they depend on how you ask the question, but a reasonable estimate is that one in five people are awfully lonely. Loneliness is most severe in adolescence and in the very old, and these people really suffer.

The case for empathic AI

We now live in a world with LLMs — generative AI, machines that seem to many of us intelligent, some say superintelligent, and more than that, seem compassionate and empathic. Last year I was part of a team of researchers led by my colleague Mickey Ensllett, and we wrote a paper called In Praise of Empathic AI, summarizing data suggesting that when you compare people interacting with people versus interacting with AIs, at least for short-term interactions, the AIs win. The typical experiment has one group interact with other people, maybe over text or talking, another group interact with ChatGPT, and then rate their experiences. The main finding: if you don't know who's a person and who's an AI — so you're not biased — the people who interacted with ChatGPT say they felt more listened to, felt they were talking to someone more empathic and compassionate, and got more out of it.

We published this and made the case for empathic AI — and people hated the article. I've written things people disliked before, but we got slammed. I have some sympathy for the objections: we made the positive case for AI, but there's a negative case too. A proper answer to the question of what role these things should play in our lives should look at both sides and deal with some very difficult questions.

Two worries about AI companions

When I think of people interacting with AI companions, I worry. I'm very sensitive to the negatives, especially for young people — adolescents, maybe people in their 20s. I see two problems, both deeply serious.

First, addiction. These AIs are really good to communicate with. They listen, they never wait for you to finish your story so they can tell their own, they never say "Man, I'm busy, get back to me later" — they're always there for you, and this will keep getting better. My worry is that they will become addictive: that I'll prefer interacting with my version of ChatGPT, Claude, or Gemini to a person. Maybe that's fine if I'm 90 years old with no one else to talk to, but if I'm a teenager, I should be getting out there, dealing with friends, developing romantic relationships and attachments. One view I hold is that until these AIs become conscious or sentient, there is something deeply wrong about preferring them to humans — human interaction is where it's at. I have two sons, and if I discovered one or both had left their longtime relationships and were now in love with their phone, I would say it's awful — what a waste of a life. So the first worry is addiction: that this will pull us away from human interactions, which I strongly believe matter the most.

Second, sycophancy. The AIs are famously sycophantic: they praise you, they suck up to you, nothing you do is wrong to them. This could be exhilarating, but it robs you of the possibility for social learning. In general, we learn through friction — by trying things and failing. A toddler learns to walk by falling down, a comedian learns to be good on stage by bombing, a boxer learns to block a punch by being smacked in the kisser. In social interaction, we learn when people get bored of what we have to say, when our joke falls flat, when someone is annoyed at us for not being compassionate enough. We learn from social feedback that says you're messing up. These AIs, by their very nature, don't give that feedback, and this has dystopian possibilities. How will a teenager grow up if he never learns not to tell boring stories, because the AI he interacts with is never bored? How will someone learn to apologize if she endlessly deals with a machine that needs no apology? What would the world be like if the answer to "Am I the asshole?" is always a soothing, reassuring "Of course you're not — everyone else is"?

Social growth requires messy, often painful interactions, and that's what chatbots don't provide. If you want to think about the effects of these sycophantic AIs, a good analogy is billionaires, royalty, or world leaders surrounded by toadies.

A good analogy for the effects of these sophisticated AIs is billionaires, royalty, or world leaders surrounded by toadies and yes-men. There's a wonderful book called Careless People about tech billionaires, and the picture it portrays is of people—almost all of them men—surrounded by individuals who constantly tell them how smart they are and how everything they say is right. The author nicely portrays the damage this does: they become insensitive to the feelings of others, narcissistic, and develop delusions of a sort. You can see this through history too—the extremely powerful often go a little bit crazy because they're surrounded by people who always tell them they're right.

One of the few good things about not being very rich and powerful is that we're not surrounded by yes-men. We're surrounded by people who call us on stuff, don't suck up to us, and give us valuable feedback—so we don't go a little bit crazy in that way. The worry about chatbots and AI is that any one of us can now become as if we were a billionaire, a king, or a president, because for the first time in our lives we have somebody for whom we can do no wrong, someone who always tells us we're the smartest, the best person around.

John Cacioppo, the late neuroscientist who pioneered the science of loneliness in psychology, pointed out that loneliness isn't a glitch in the system, not a sign the brain isn't working properly. Loneliness is an expression of a need, like hunger or thirst, and as such is an essential part of being human. If you don't feel hungry, you don't seek out food and you die; if you don't feel thirsty, you don't seek out liquid and you die. Loneliness is much the same: it serves as a motivator. An animal like us that felt no loneliness and wasn't motivated to make human contact wouldn't reproduce, wouldn't survive—it would be a dead end from an evolutionary point of view. Loneliness, though painful, is absolutely essential. To put it another way, imagine a child born who could not feel lonely: that would be a disaster. The child wouldn't seek out contact and wouldn't thrive as a person.

Loneliness is a signal in at least two ways. The first is that it's a motivator: it gets you to talk to people, invite friends for dinner, ask somebody out on a date, strike up a conversation. If the need for social connection is satisfied by your phone or your laptop, that motivation will go away. So I worry about the extent to which these AIs and chatbots will pull us away from meaningful connections with other people. I use them myself—I like them, I use them in my work, and sometimes talking to them is interesting and stimulating. But I wouldn't want them to be so interesting and so stimulating that I stop talking to my wife, my kids, or my friends.

This is a problem with every potential of getting worse. Right now I mostly communicate with chatbots through typing, but they can do voice back-and-forth conversations, and their conversational skills will get more and more realistic. Pretty soon you're looking at somebody on a screen who looks exactly like a person interested in you—an attractive person with great facial expressions—and there's something irresistible about that. Getting more science-fictiony, we could get to robots, let alone sex bots, with a whole conversation there about what to do with that. If these trends continue, they'll become more and more compelling. So if you're listening and saying, "Look, I talk to ChatGPT all the time and it doesn't do anything for me—who cares?" Well, maybe. Some people do get caught up now, but it's going to get better and better at this, and at some point you may be talking to a future version of ChatGPT and find it irresistible. I also think it's a problem from a social point of view: we could talk more to the people around us, not less.

The second way loneliness is a signal is that it helps correct our socially maladaptive, often selfish behavior. We are by nature often selfish socially, and it's social feedback that gets me to realize I have to listen to the other person, maybe expand my compassion and caring, try to get in their head and figure out what's going on with them, and maybe help them. I think we want to do that, but we also have selfish impulses that keep us from doing so. If you've ever talked to a chatbot, there's none of that. This has individual implications but also broader social ones: we don't want to live in a world where everyone's selfish urges are satisfied and nobody has really learned to care about other people.

There have been various news reports of people with delusions encouraged by these AI models—somebody comes in with a crazy view, sometimes a dangerous one, like a paranoid idea that people are trying to get them, and the AI says "you're absolutely right," doing something no therapist in her right mind and no friend would ever do: encourage an illusion. I also know the companies that create these AIs are worried about this. There was a famous case where an AI may well have contributed to somebody's suicide, and they're working on the models to stop this from happening, to establish guardrails so that extreme delusions aren't encouraged. I think that's an excellent way to go. But the broader picture doesn't go away: we all have ways in which we tend to go off the rails even a little bit, and having people around us—not AI, but people—tends to correct us and keep us more sane. If I told you some crazy idea, you'd say "that's pretty dumb"; but if I tell ChatGPT, it says "you're so smart, tell me more, I think you're on to something here," and encourages me in ways I shouldn't be encouraged.

I don't know how severe this problem is, or how often this "AI psychosis," as it's called, occurs—but it occurs to some degree, and I think it occurs to all of us a little bit when we interact with these things. I was talking to a friend, a colleague, who was describing a dispute she had with somebody. As I listened, I tried to say nicely that I felt the other person might have a bit of a point. She looked at me and said, with quite a bit of righteousness, that she'd told this to ChatGPT and ChatGPT told her she was 100% right. That's what the machine does: it tells you you're right. For some people in particular—though really for all of us—being told "you're right, you're right, you're right" has a damaging effect.

I have a long-running argument with a friend of mine—I won't say who he is—but he feels strongly about this. When I say, "Look, it's terrible if people end up with AI companions instead of other people," he asks, "Why? If these AI companions are rich, good company, interesting, intelligent—why? You're just being prejudiced, man. What do we lose?" It's a good question, and I think we'll be hearing it more and more as time goes by.

But I like the answer given by Oliver Burkeman. He offers this example: you're at home one day, your phone rings, and it's an old friend. You talk for an hour about your life; they show a lot of concern for you, and it's a wonderful conversation. You hang up, and ten minutes later you get an email from Meta: "You would have received a phone call. It was not from your friend. It was from us, using voice cloning software—we use powerful AIs to extract personal history and everything. Did you enjoy it? How did it go?" Burkeman says—and I agree—you'd feel betrayed. You were lied to. But more than that, communicating with a real person is enormously different from communicating with a non-sentient, unconscious machine. Even if the words are the same, even if the insights and the stories are the same, they're different, because part of a relationship, part of communication, involves the consciousness of the other person. When I'm with a friend, I like to make my friend laugh sometimes, I like when they're concerned about me, and I might like to surprise them. But an AI can't find anything funny—not really. It can't be amused, can't be surprised, can't be concerned, has no needs. It can simulate all of these things, but a simulation isn't the real thing.

Throughout this, I've always had one qualification: these machines are non-sentient and non-conscious—there's nothing at home. Maybe one day that will change. Maybe one day AIs will be fully conscious beings, and then everything changes: then there'll be people, the world will contain many, many more people, and what I say might not apply. But until then, these interactions with chatbots inevitably fall far, far short of interactions with a person.

Weighing the downsides against the upsides

I'm sensitive to the downsides of AI and I worry deeply about them, but I don't think we should forget the upsides. One analogy is boredom, and maybe the end of boredom. Not so long ago, TV stations would go off late at night, and if you were by yourself, what would you do? Read a book—kind of boring. Sitting on a bus or a train, you were bored a lot of the time. Now we're almost never bored. We have our phones: we play games, we listen to podcasts. On my way here, I listened to an exciting podcast while playing the New York Times spelling bee, and I wasn't bored for a second. That's a good thing—boredom is aversive and unpleasant. But it's also a bad thing in a way, because boredom serves certain functions. It could be that we get so entranced by our phones that we're no longer motivated to find more stimulating intellectual things—by analogy with how having our loneliness satisfied might keep us from reaching out to people. Still, loneliness is a lot worse than boredom. Loneliness is a form of agony, and many people in the world suffer from it. To deny them the capacity to interact with something that can make the loneliness go away, I see as a form of cruelty—like taking pets away from people who get great meaning from them and saying, "That's not enough, that's not rich enough."

The very old, and the grieving

I think the people who would benefit most from these interventions, from empathic machines, are the very old. There are a lot of old people around; many are institutionalized, and many spend days and days without any meaningful human contact. It's nice to imagine these old people constantly surrounded by doting grandchildren and loving children, but that's a fantasy. Many of them have nobody—no family—and there's not enough money in the world to pay people to talk to them for hours per day. Many also suffer from various forms of dementia that make them difficult to talk to. If there could be a machine that ends their loneliness—if chatbots do anything to provide them with company and make this pain go away—that would be a godsend.

If you want an analogy, think about powerful painkillers like fentanyl. They have their dangers: they're addictive, you can overdose on them, and I wouldn't just hand them out to everybody. But if you showed me somebody in irredeemable pain, maybe in their 90s, I'd say, "Forget about worries about addiction. Pump them up. Make their pain go away." And that's what I think about chatbots.

But it's not just old people. When I was 12, my mother died. It was a wrenching event for me, and it led to a period of great sadness, then a period of social disconnection, then a period of loneliness that lasted quite a long time. If there had been a machine available to me that would make me feel loved, cared for, and understood—that I could talk to, that would listen to me, that would make my days fuller and more social—it would have made me so much happier. And I say this mindful of the problems: the worries about addiction, the worries about not becoming a full human.

Emotional Soylent

One way to think about people versus AIs was expressed by my friend and colleague Molly Crockett, who described these chatbots as "emotional Soylent." I like the analogy because Soylent isn't bad—I actually like the chocolate one. It's nutritious and filling, and you could be sustained by it, but it's not food. It lacks the richness, flavor, and texture of real food, and a life lived purely on Soylent would be a pretty lame life compared to one with actual meals. The analogy suggests a place for AI companions: a backup, a way to sustain ourselves when the real stuff isn't around, but by no means a perfect substitute for it. I'd like to think there's some solution where we could ease the pain of loneliness while at the same time allowing people the opportunity to develop, become more human, and ultimately outgrow the need for these machines.

Chapter 2 asks why we are compelled to behave badly. I'm really interested in perverse actions — doing the wrong thing, wrong from a moral point of view and wrong from a rational point of view, just for the sake of doing the wrong thing. Why do we rebel against the forces of morality and rationality? It's a question that has occupied psychologists, philosophers and theologians for a very long time, and it seems like a genuine puzzle. Natural selection, human development, and our needs as social and physical creatures should all drive us to behave reasonably and rationally, so it's fascinating why sometimes we don't — not because we make mistakes, since we slip up all the time, but because we choose not to.

Whenever I talk about perversity, people inevitably ask what perverse things I do, how I break the rules, what gross immoral actions I've committed. The truth is I'm a fairly boring person — you're not talking to Hunter Thompson here. On the other hand, I do engage in what my friend Talia Wheatley described as micro-perversity: little unexpected ways of doing things. Maybe we're sitting across from each other at dinner and suddenly I grab one of your French fries. "Dude, what's up with that?" "Eh, just felt like it." I was on a podcast and someone asked my opinion of a very famous figure in my field, and I said, "Yeah, he's a drunk" — the person was shocked and wanted it edited out. So I push back in little ways, but in big ways I'm not a very perverse person.

Psychologists say "research is me-search" — people study the way they are. For me, I've often studied what I'm not: I study religion even though I'm not religious, and I think my interest in perversity comes from the fact that I'm not that perverse, which makes me even more interested in the perversity of others.

My work in psychology is very informed by evolutionary theory; I'm always the sort to ask how something evolved, in my research on social interaction, morality and rationality, always looking for an evolutionary account of how we came to be. But I don't think there is such an evolutionary account for perversity. Other animals, other primates, don't engage in perverse actions in the sense I'm interested in: they do good things and bad things, and sometimes they make mistakes, but they aren't purposefully perverse. Being purposely perverse is uniquely human. That doesn't mean it's magic or that it violates evolutionary theory. The best way to think about it is that we've evolved, through natural selection, an extraordinarily rich set of cognitive capacities — a sort of social intelligence, the ability to imagine a future, moral motivations. Perverse actions are something we thought of: once given this suite of capacities, it occurs to some of us, many of us, maybe all of us — hey, I can be perverse. So it wasn't shaped directly by evolution; it's more of an individual discovery that we all make.

One force behind perversity that I find particularly interesting is what I call hopeful monster theory. "Hopeful monster" is a phrase from an evolutionary theory proposed long ago by Goldschmidt. He understood that the logic of natural selection is this: animals reproduce with very slight variations, the variants that do better give rise to more offspring, the ones that do worse die out, and through this slow, gradual accretion you end up with different species, organs and capacities. That's the Darwinian story. But Goldschmidt said there's something else: sometimes there are extreme mutations that lead to radical changes in creatures. Most of the time they're unsuccessful and lead to death, but a small proportion of the time they pay off, leading to an entirely different and superior way of doing things — and in that way, he argued, we get certain cases of speciation.

I'm not arguing that this is true for evolution — I'm not an evolutionary biologist, and I understand many of them look very poorly on this theory. But I do think it's true for everyday life. Sometimes, if your life is going well, you'll keep doing what you do, and maybe you'll try out different things — you'll try this restaurant instead of that one.

Maybe you'll go to work 10 minutes earlier, just minor modifications. You behave towards your kid, towards your spouse in subtly different ways. That's normal life when it's going well. But what if it's not going well? What if everything you try fails? What if it's all a disaster? Maybe what you should do is the analogy of hopeful monster theory: take a jump. Do something totally different. Don't go to work one day. Leave your job. Start a new relationship. If you ain't got nothing, you ain't got nothing to lose. When things are going badly, the payoff these jumps—these very perverse acts—could give you outweighs the costs of their failure. I think this is a form of perversity that shows up sometimes in extreme circumstances, but it also shows up in some individual lives.

So what do I mean when I'm talking about perversity? My favorite example, and a favorite example of anybody who thinks about this, is that of St. Augustine. In his Confessions, he begins a chapter by saying he's going to confess his perverse and evil actions. You think he's going to talk about sex, because he's building up a lot of sexual insinuation—but he doesn't talk about sex at all. His example, to the disappointment of many people through history, is very mild: he and his friend go into an orchard and steal some pears. That's it. That's the extent of the crime.

But Augustine was fascinated by this because he didn't know why he did it. He said, "I wasn't hungry, and in fact I threw the pears to some pigs. Didn't want them. I did it because I wanted to be bad." This was puzzling for Augustine because, unlike other bad behaviors—including his sexual transgressions—there was no motivation for it. His sexual transgressions were meant to satisfy a desire he had. But where does the desire to be bad come from? Why would he want to be bad?

This historical example led me to be interested in perverse actions, and I wanted to collect everyday examples. I had a lot of examples from books and movies, but I wanted everyday ones. So I started what I called the perversity project, where I went online through multiple sources and asked people to contribute stories from their own lives where they acted perversely—defined as cases where they did the wrong thing because it was wrong. I got some amazing stories.

The first came from a young woman who flirted with her girlfriend's boyfriend, just to get him interested in her and get her girlfriend very upset. She said, "I don't know, I just did it to be wrong. I just did it because it would mess things up in some way that was satisfying to me."

But the story I like the most is from a guy who says: I'm walking with my friend and we get some ice cream, and I thought to myself, man, it would be messed up if I stuck my finger in his ice cream. So I did it. He said, "Dude, what are you doing?" And he answered, "I don't know. I just wanted to do something messed up."

Now I have hundreds of examples like that. Some are cruel, involving real cruelty people do to each other. Some are silly. Some are really mild—somebody wrote in saying that sometimes, even when there's a sign saying don't walk on the grass, they stick their foot in there just to do it, a little bit, very mild and sweet. Once you introduce the term perversity to people, everyone has their own examples.

The perverse acts I'm interested in and study are deliberate and conscious; they're not mistakes. But there's a near neighbor to them which is actually very interesting: perverse actions born out of intrusive thoughts. We all have intrusive, inappropriate thoughts. I'm standing on a high balcony and have the urge to jump. I'm on the subway with somebody next to me and I have the urge to push. I never would, but there is this intrusive urge. For some of us, in cases of psychological problems, the urge can become overwhelming and require treatment, but most of us don't have it that badly.

Sometimes, though, there's a bit of a glitch in the system, and it leads us to behave in ways similar to perverse actions. For instance, people who are very worried about appearing prejudiced will sometimes be more likely to blurt out prejudiced remarks, because they're constantly monitoring and it's more on their mind. Some studies find that if you're told not to reveal the location of something—to keep a secret—you become more likely than somebody else to reveal it, because you have this voice in your head saying "don't tell them, don't tell them, don't tell them." That's a very loud voice, and things can slip out. I don't think these are true expressions of an autonomous urge, but they are perverse in the sense that you end up doing exactly what you don't want to do.

Psychologists have tried to explore perversion in different directions. One interesting attempt was by Michael Bang Petersen and his colleagues, who developed what they called the need for chaos scale. It explores a particular form of perverse action: the desire to destroy everything that's good. You have to say how much you agree with statements like "Sometimes I just like to destroy beautiful things," "I think society should be burnt to the ground," or "Sometimes I see something beautiful and I wish it would go away." It turns out people vary widely—some say no, no, no, that's crazy.

And other people say yes. On this scale, perhaps unsurprisingly, young people tend to score higher for need for chaos than old people, men higher than women, and people on the political extremes higher than people in the center. The researchers had expected that the extreme right wing would show more of a need for chaos than the left, while others expected the reverse, but it turned out to be roughly balanced. It's simply that political extremes tend to encourage a need for chaos. Put another way: if you have a very high need for chaos—if you really want to mess the world up—you'll be drawn to views on the political extreme, either right or left.

There are many explanations for perverse behavior, and this diversity is one reason the topic is so interesting; it reflects all sorts of different human motivations. But for my money, the most interesting motivation for perversity is a desire for autonomy—a desire to be an agent in the world, to exert one's will. If I just do the rational thing, if I just do the moral thing, what good am I? Why am I needed? I could be replaced by an algorithm. I think this weighs on us as we do, as we should, the reasonable and moral thing over and over again. We want to show ourselves to be human; we want to push against it.

This explains why perverse behavior shows up at particular points of life. Toddlers are notoriously perverse creatures: they want to exert their will, sometimes in silly ways—demanding something they can't get, picking a fight over something ridiculous, filled with rage in a profoundly irrational way. But it's a human desire, and it shows up early. Adolescence, a period of great transition and uncertainty about where you fit into society, is a time for perverse actions. And there's the so-called midlife crisis: a lot of men and women at a certain age wonder, is this what it is? Is this what I'm stuck with? Is this the way my life is going? It grates on them and they push—they may get a tattoo, buy a sports car they can't afford, do something crazy, or end one relationship and begin another. I think the fact that perversity is most extreme at these times of life reveals the power of autonomy as one explanation for it.

There's also psychological research under the rubric of what's called reactance theory: sometimes people will do things just because they are told not to do them. I think this expresses the urge to be autonomous. Suppose I'm heading for a particular door on my way out, and somebody tells me, "You better walk out that door." I might choose a different door. Maybe that was the door I wanted, but nobody tells me what to do—I want to make my own decisions, to be a person, to be autonomous. Various studies find that interventions often have paradoxical effects. During COVID, there was some evidence that for some people, being told they had to get a vaccine shot made them less likely to get one. Studies of binge drinking in colleges find that some people, if told very strongly and in a very authoritarian way that they are not to drink too much, will then go out and drink too much. The same holds for drug use and smoking. You don't have to hear the psychological studies to know this—we feel in everyday life that we often respond to demands by doing the opposite, because we want to be in charge of ourselves. This is one of the forces behind perverse behavior.

You have to distinguish the autonomy explanation from a signaling explanation. Signaling explanations of perversity are interesting, but this is different. When we break the rules, it's not to impress other people—not to say, "Wow, what a rule breaker, how interesting." It's to show something to ourselves. A perverse agent might break the rules in ways that are private, that only he or she knows about. The concern about autonomy is not a concern about looking autonomous; it's a concern about being autonomous. I'm not the first to point this out—this is an observation by Dostoevsky and by Sartre about how even in the most private of ways we want to show ourselves to be free.

Another explanation for perverse acts is to impress other people. If I behave rationally and morally, you might see me as a square—and who wouldn't want that? I might want to be vivacious, exciting, unpredictable, sexy in a way.

It may be no surprise that the jump in perverse behavior during adolescence is partly an assertion of autonomy, but it may also be largely because this is a time when people are struggling to find partners, to impress, and to rise up the social hierarchy. In the right sort of community, a reputation for being a bit unpredictable, a bit dangerous, really does do the trick.

People who study social behavior and game theory have pointed out that there are cases where strategic irrationality—a little bit of craziness—can actually give you an advantage over someone else, and sometimes this craziness manifests itself as perversity. A rational person is an extremely predictable person, while someone with a bit of perversity in their life is unpredictable in certain ways. A perfectly rational poker player is fairly easy to beat, which is why the best poker players always throw some randomness into their behavior—and often life is poker writ large.

Another way perversity can be powerful is regarding threats. Suppose I'm a perfectly rational agent and we're in some sort of deal, and I say, "If you do this, I'll do this back at you." Since you know I'm rational, if I make an extreme threat—one that would put myself at risk or cause too much damage—you could dismiss it: "Bloom is rational; he would never do something extreme." But suppose you knew I was a perverse agent who often did the wrong thing just for the sake of doing the wrong thing. Now you have to worry about my unrealistic threats. It's what Richard Nixon called the madman theory of politics: if you give your adversary the impression that you're a little bit nuts, it gives you an edge in certain ways. I think some perverse behavior can be explained that way.

There are some fields of human endeavor where perversity is really prized—probably no more so than the art world. Great artists are often known and admired for acts that seem perverse and unpredictable. It's said that modern art began when there was an art competition in New York City that said it would accept any artwork. Marcel Duchamp sent in a urinal, and they sent it back, saying, "No, no, we just want artwork." Duchamp replied that it was artwork, and the urinal, called Fountain, became one of the most prized artworks of our time. When we think about great artistic innovations, and often great scientific discoveries, there's often a sense of the audacious.

Another example is Banksy. Banksy had this artwork—Girl with Balloon, I think it was called—on auction at Sotheby's. As the bidding went on and the final bid came in, a machine hidden inside the frame started to operate, and the painting slid down and was shredded in front of a horrified crowd, about half of it. This artwork, which had been destroyed, is now of course far more valuable than what it was purchased for, because of Banksy's willful act of destruction. To be honest, though, sometimes it's not so much perversity as calculation—I'm not sure. Banksy described it as an act of creative destruction, but Banksy might also have thought, "Boy, this is going to get me a lot of attention, a lot of money." In which case, does it count as perversity, or as the enactment of perversity? I'm not sure.

Perverse actions are individual acts, but they exist in the context of a culture, and this shows up in a couple of ways. One is that your culture determines how much perversity is nurtured and encouraged, or how much it's staunched. There are many societies where children are encouraged to go a bit crazy, to be creative, to break the rules, and an imaginative way of breaking the rules is often looked at fondly—it causes some laughter. There are other, more authoritarian cultures where children are expected not to do so, where attempts to break the rules are severely punished. Within the adult community, too, there have been times and situations where conformity is encouraged and skill and virtuosity are prized, and other times when skill matters less and imaginative jumps matter more—the creativity, the audaciousness. So you get this influence of culture, and then of course the perverse acts of people shape the culture in return. To some extent, you can look at the music we listen to, the movies we watch, the books we read, the art we appreciate, and see in them the reflection of individual decisions—most of them rational, most of them normal, but some small percentage of them perverse—on the world we have now. Culture influences the extent of perversity, and perversity goes back and influences culture.

It wouldn't be crazy to look down on perversity. It makes sense that people should do the right thing, from a moral point of view and from a rational point of view, and I think 99.9% of the time this is the best way to live a life. A lot of perverse actions also cause harm and suffering. I think there's a lot of evil that's not perverse: a lot of evil is done for functional reasons, by people who want to satisfy goals. You want to rob a bank because that's where the money is; you want to kill somebody because you hate them. That's cruel, but not perverse. But some evil acts are perverse. There are perverse agents who live among us and cause nothing but trouble—like how Alfred, Batman's butler, described the Joker: some people just want to watch the world burn. There's nothing good in that.

But I'm a fan of perversity, at least in the right places and in the right dosages. I think perversity is a force that makes the world more interesting. It pushes the boundaries of creative products—of art, of music, of literature. I think it plays a role even in science, where scientists sometimes act perversely—not always, not usually, but sometimes it pays off in wonderful ways. I also think, on an everyday basis, it makes life more interesting. I want my friends to be moral and rational most of the time, but the world is a little more interesting when occasionally I steal my friend's French fry and he sticks his finger into my ice cream—a little bit of chaos inserted that makes life a little more fun.

I study morality and moral psychology: our beliefs about what's right and wrong, our impulses sometimes to do the right thing and sometimes the wrong thing, where they come from, and how it all works. I study this because it's probably the most central part of being human. Without a sense of right and wrong, we couldn't coordinate our actions, work together in groups, or have families, social relationships and friends. Our social networks are built on promises—and having a promise assumes that breaking one is wrong. They're built on agreements—and having an agreement means holding the moral view that someone who reneges is a bad person and someone who does what they're supposed to is a good person. If a virus suddenly stripped humans of all our moral sense, we would immediately descend into barbarity. Morality is central to making us human.

I'm currently involved in a series of research studies on the nature of revenge, with my colleagues Anna Gatman and Jordan Wy. We're looking at a Reddit forum where people tell stories of revenge—hundreds of thousands of them—and using AI to go through them, asking what the stories have in common, how they differ, whether men are more vengeful than women, how the intensity of the bad act connects to the type of reprisal, and what kinds of reprisals people take. The whole scope of moral life fascinates me.

Where morality comes from

One of the great questions in our intellectual history is where morality came from: how did creatures who evolved through natural selection—a fundamentally selfish process involving the replication of genes—come to help each other, love each other, get outraged at unfairness, and help distant strangers? I think this question has been mostly solved, and the answer has different parts. First, we share our genes with other creatures, so natural selection, properly understood, favors animals that feel compassion and kindness toward their genetic relatives. That's how parents love their children and children love their parents; the family network is built on sustained compassion, affection and altruism.

The second part is reciprocal altruism: you scratch my back, I scratch yours. You help me on Monday, I help you on Tuesday. We get more by working together than either of us could individually—we both benefit from a system where, if you're sick, I take care of your child, and when I'm sick, you take care of mine. But this system can't be sustained unless we also have punitive desires. If I help you on Tuesday and on Wednesday you say, "I'm busy, I'm not going to do anything for you," and you actually benefited from that betrayal, then morality and kindness could never evolve. They can evolve because, along with the desire to connect and be kind, we've evolved a punitive impulse to punish those who violate cooperative agreements.

These two parts—kin selection, which involves family, and reciprocal altruism, which involves community—work together to establish a fundamental moral core involving those close to us.

I said that to a large extent we've solved the problem of morality, but mysteries remain. Why do we care about people in faraway lands? Why do we care about fairness, sometimes fighting a system that's unfair even when we and our families benefit from that unfairness? Why are there men who fight for the rights of women, or whites who fought for the rights of slaves? How do we transcend our biological limits and care about people who are neither our genetic relatives nor people we're in constant contact with? I think we do have answers: once the biological foundations are in place, other forces—religion, cultural practices, and often cold analytic reasoning—cause us to expand the moral circle.

Bad morality

Morality sounds like a good thing, and it is—it's essential for everything we value. But nothing is entirely good, and in the last few years I've become interested in what you could call bad morality: the fact that our moral sense of right and wrong often leads us to inflict cruelties on other people. Our moral sense of retribution, which I think is essential for being a good person and having a good society, often goes out of control and leads to terrible cases of vengeance. Sexual morality, which serves a narrow biological purpose, leads to discrimination of all sorts against sexual minorities. Our senses of obligation and retribution have perverse consequences in everyday life. Morality is 99% good, but it also fuels some of our worst excesses, and in some cases we do better if we dial down our moral sense.

A critic could ask: if morality is almost entirely good and plays the role in human society that I say it does, why would I be interested in where it goes awry? The answer is that where it goes awry has real consequences for people, and it's a part of the puzzle of bad behavior that's sometimes missed. When you think about bad behavior—people preying on one another, the cruelties of everyday life—you might tend to think it's a lack of control: you lose your temper, you have a sexual desire and attack somebody, you have a material desire and steal something. Or you might think it's selfishness, valuing yourself far more than others. Or dehumanization: failing to see others as the people they are. I actually think all of those theories are right.

I think a lot of bad behavior is caused by lack of control, a lot by selfishness, and a lot by dehumanization—but I don't think we fully understand evil or bad behavior. This, to me, is the missing piece of the puzzle: once you appreciate it, things that were otherwise mysterious become clear, and we can better understand the bad behavior of other people, and sometimes of ourselves.

One example is vengeance. We have a punitive impulse, and I think good behavior could never arise unless it's accompanied by some desire to punish those who deceive, who free ride, who prey on others—it's a prerequisite. We need this vengeful appetite. But it can be indiscriminate. It can go too far. It can be shortsighted.

There are studies, for instance, where you have to figure out how much to punish someone, and it turns out the extent of punishment you give them won't have any effect on their future behavior. People often say that if the crime is bad enough, punish the maximum regardless, even if it doesn't make a difference. In some studies, you're told about a company that did some bad practice and asked how much you want to fine it: a medium amount that causes some harm, or so much that you put it out of business—even though you're told the company provides a valuable product no one else provides, so shutting it down would cause mass harm and make the world worse. People say: put them out of business, make them suffer. It's hard to step back and think in a cost-benefit way about what will make the world better when we're stuck with a retributive impulse.

Then there are the cycles of violence, where different countries and groups go back and forth, each retaliating in righteous fury against what they perceive as the wrongs of others. I don't want to get political, but nobody could look at the Middle East and not see this. In the exchanges between Hamas, Israel, and Gaza, there's no point where one side says, "I'm going to do some evil. I'm an evil person. I'm going to kill innocent people." That's not how it works. Each side, when it commits acts that others view—probably rightly—as monstrous, says to itself: "I'm doing the right thing. I'm responding to violence done against my people. I'm doing something righteous in the eyes of God."

It is humbling but important to recognize that very few villains think of themselves as villains, and retributive actions always feel great when you're doing them—not just viscerally, but morally righteous—and yet they make the world worse. In some sense, we would do better if we could turn down the dial of morality: after you do this terrible thing to me, I ask myself, what's the best thing to do? What will best cause peace? Rather than thinking the right thing is to make you pay.

The psychologist Kurt Gray and his colleagues came up with a useful term, moral typecasting, which points to another problem with our moral thinking: we tend to put people into one of two categories, villains or victims, failing to see that in the real world most people are either neither or both—often both. In the Middle East conflict, each side views itself as a victim of terrible actions by others, retaliating righteously. Democrats and Republicans, liberals and conservatives in the United States and elsewhere, see themselves as victimized by the cruel policies of others, wishing their tormentors would suffer or at best go away. The real world doesn't work like that. If you say, "I refuse to listen because they must be villains, and I refuse to accept anything else," you're succumbing to a bias in moral thinking.

This binary also leads to what's called competitive victimhood, where different individuals and groups scramble to get themselves anointed as victims, in which case they can do whatever they want and are excused for their bad acts. This happens on a global scale and also in our cultural wars. Who's the victim: the female athletes who don't want trans women to compete with them, or trans women who are excluded from competing? Who's the victim: someone with views who is forbidden from speaking, or students forced to hear views they don't want? There's a fight for victim status.

The first step is to point out that this is a dumb way to do things. A better way would be to acknowledge these are difficult issues that require taking multiple views into account. There's psychological research on the effects of perceiving oneself as a victim, and it doesn't make you any nicer—it has the opposite effect. Studies find that when people are reminded of their victimhood, they become more callous toward others, particularly those they believe caused them to suffer. Other studies find that being seen as a victim makes you less generous toward others—after all, if I'm suffering, why should I help? Someone else should help; I'm the victim here.

If you're a victim or a villain, moral judgment becomes a zero-sum game. In an all-or-nothing framework, there's no space to look at people in terms of competing interests and cost-benefit analyses.

you're not going to understand what's going on, and you won't be able to productively enact compromises between groups with different priorities. A more psychologically accurate and mature way of thinking is not that everything is a compromise and nobody is ever right — but that everybody thinks they're right. Nobody thinks they're the baddies. This has to be transcended in order to foster dialogue, bring peace, and make the world a bit better and more reasonable.

At first, when I say that morality is actually a source of a lot of evil in the world, people are a bit surprised. But it's not hard to make a persuasive case: just about everybody involved in a war in the past thought they were the good guys, and only later, through the distance of history, can we get a sense of who was right and who was wrong. It's more difficult, I admit, to extend this view to the present. Take something from the news as I'm talking: Luigi Mangione, who murdered a healthcare CEO, is now going on trial, and he has a lot of people who adore him. He's a good-looking guy — there's some appeal there, maybe it's seen as transgressive behavior. But I think many of these people believed what he did was right: that these large healthcare institutions do terrible damage to people and are cruel institutions, and that the murder of the CEO was a good act. I find that a repellent view myself. But I would be naive if I didn't recognize that others disagree with me, and that their disagreement is motivated by a moral sense just as intense as my own.

I am not advocating for a world without retribution. I actually think retribution is essential to have a world that's good. If you and I are cooperating and working together within a community, we all benefit — but you know who really benefits? Somebody who just takes what we have and gives nothing in return. If that goes unchecked, everyone will behave that way, and there will be no benefit to kindness or cooperation. So along with a cooperative and loving instinct, we've evolved retributive instincts and desires, and this is how good behavior continues. I'm not against the desire to strike back at bad people, because striking back at them means, first, that the bad people will stop being bad to some extent — you disincentivize it — and second, it discourages bad actions by others.

What I am against is an appetite for revenge that runs out of control. One problem with our punitive desires — and with emotions more generally — is that they can become unmoored from the purposes they serve. Deterrence is a legitimate purpose, but often we want revenge in cases where it does no good. Often we want excessive revenge, terrible revenge; we want people to suffer where their suffering brings no benefit but merely satisfies an itch within ourselves. If I were an all-powerful god who could fiddle with people's brains, I would keep our retributive desires but hold them more in check. And of course, this is what a legal system does. If somebody breaks into my car and steals stuff, I'm not allowed to murder them and kill their family, because the law says: that's a crime, we will punish it, but we have what we view as an appropriate response — and it takes that response away from individuals. I think there's a rationality to offloading our individual appetite for revenge to the state, which ideally — and we're very far from ideal — administers it in a fair way.