Kazuo Ishiguro on emotional truth in sci-fi and in our AI reality

A novelist meets science

0:00Hey, it's Flora, and you're listening to Science Friday. The movie Clara and the Sun comes out in theaters today. It's based on the 2021 novel of the same name by Kazuo Ishiguro. The story is set in the future and is narrated by Clara, a robotic, artificial friend whose job is to be a companion to a sick teenage girl. Novelist Kazuo Ishiguro has spent his career writing about the human condition, often in relation to changing technology and scientific capabilities. His other novels include Never Let Me Go and Remains of the Day. He won the Nobel Prize for Literature in 2017.

0:45Kazuo, I'm so grateful that you're joining us. I have been reading your work my whole adult life, and it really has given me so much. So thank you for making time to come on Science Friday. Thank you. It's terrific to be here. There's a lot of science in your books, you know, not just robotics or AI, but gene editing and memory. I'd love to start with just hearing a little bit about your personal relationship with science. Well, it's not a great one, actually. My father was a scientist, and that's partly why, you know, a Japanese family was living in Britain.

1:24My father was working for the British government as a research scientist, and he was a kind of an inventor. He was trying to create a machine that could predict tidal waves and so on. The machine that he created is now in the Science Museum in London, because it is such an interesting piece of innovation that sits right between the analog era and the digital one. So this is all happening, more or less at home, because he worked at home a lot. He'll go to his laboratory.

1:56But one of my great regrets is that I didn't pay very much attention to what he was doing. The trouble in, certainly in Britain, I don't know how it is in America, is that there is almost like a cultural divide between science and arts. And I kind of felt I was an artistic person, and I was interested in music and books and things. And so I gave myself license to completely ignore what my father was doing. And I wish now that I had paid much more attention.

2:29As I get older, I mean, I feel there's this kind of huge gap in my, not just in my knowledge, but in the way I think, that wouldn't be there if I'd trained myself to think more in that kind of scientific way. When did you reacquaint yourself with science? Oh, I've never properly reacquainted myself with science, but, you know, I have some scientist friends or, you know, people who work in the tech world.

3:01And, you know, talking to them is very illuminating.

3:06And one of the interesting surprises for me is that the two worlds aren't that different. You know, I always imagine that things are much more clean cut in science. You know, people reach conclusions and consensus, and that was it, and then they moved on to do the next thing. You know, when I get to know people in science, they're having rows and disagreements, just like people in the cultural world. There's fierce rivalries and accusations being thrown around all the time.

3:40And in a way, it is comforting that science doesn't progress in this kind of clean cut way, that it has the mess of human debate and questioning about it. I mean, it's an inquiry. And people seem to acknowledge in the world of science, just as much as in other worlds, that, you know, reasoning can be motivated. And just because it's scientific reasoning, I mean, it doesn't mean to say it's not motivated. And so I'm kind of reassured in a funny kind of way by the messiness of the scientific world.

4:17You know, I just had a conversation with a scientist yesterday, and he said there's this popular misconception about science that it is about answers, when in fact, science is really about questions. I'm sure that's true. But, I mean, you know, there's no point in asking a question if you're not looking for an answer. That's true, too. And so in the end, yes, we all aspire towards answers.

Scientific truth versus emotional truth

4:40But it's been interesting for me, particularly, you know, in recent years, at times there seems to be a kind of a standoff, if you like, between the kind of truth that, you know, people like me or, you know, filmmakers like Taika Waititi, who made the movie of Clara and the Sun, there's a kind of a truth that we're reaching for when we do our work. Which is kind of, I think it's different in kind to the kind of truth that scientists are arguing about and trying to reach.

5:16And in recent years, I feel that the competition between these two ideas of truth seems to be more visible, particularly during the pandemic, where all of us, just at every level, we are so hungry for answers to questions from the world of science. We wanted to know which vaccines were going to work or whether we're going to have a vaccine, how, you know, what's happening with this virus? Is it going to get worse? We were so tuned in to the scientific community, wanting answers from them.

5:51And on the other hand, we seem to drift right to the other end of the scale about what you might loosely call emotional truth. You know, we wanted to believe what we wanted to believe, you know, we had conspiracy theories, you know, maybe vaccines were part of a conspiracy theory. It wasn't just about the virus. I mean, it just seemed to be in the air that people felt that if they felt something in a strong enough way, that was the truth in some kind of way, you know.

6:26And now we're living in this time of people debating about AI and to a lesser extent, you know, gene editing.

6:39But I see that same conflict. People want to believe what they want to believe. And just having the presence of AI in the world seem to arouse very strong emotions, not just toward AI itself or science itself, but it seems to be doing things to our relationships with each other, you know, in the family, you know, how we look at each other, how we feel about each other, even the fundamental things we assume about ourselves.

7:11Is that happening for you with AI? Are you looking at yourself differently or your family differently?

7:20It's hard to say. I mean, we like to think there are certain kind of feelings and emotions that are eternal or universal to all human civilizations. And we all care about our children. We have a certain bond amongst family members. We look for spouses or partners.

7:41And we have this thing called love. You know, we fear death. I mean, some things we've always assumed are eternal almost, you know, part of our human condition. But they've always taken place within quite powerful contexts, I would say. So, you know, in a more religious age, what you believed about God and what happened after death, I mean, that had a big impact on the way you related to people. You know, sometimes you could break off a relationship with somebody you loved because it went against some kind of belief system that you live within.

8:20And I think we are now moving into an age when we are, a lot of our older fundamental beliefs are being challenged by things like AI.

8:30And I think inevitably, perhaps almost incrementally without realizing it, I think we may start to relate to each other in a different way.

Writing human connection in Clara and the Sun

8:41And so, you know, when I was writing Clara and the Sun, and I think very much in the movie, we were looking at family relations and emotions like love. I mean, this story isn't about, it's not trying to predict where we might go if the technology and the science gets out of control. It's not necessarily suggesting that AI or gene editing is a menace, although, you know, it definitely acknowledges the darker side of those things.

9:15But in our story, we focus on what happens to human beings and their assumptions about what another human being is. And can you make sense of an emotion that you label as love? If the possibility is raised that, you know, you can find out almost everything, excavate every kind of piece of information about the person, the object of your love, and maybe even transfer it or, you know, preserve it somewhere so that you can transfer it.

9:47I mean, if somebody is just a bundle of impulses and emotions that you can more or less predict, maybe you relate to them in a different way. You know, you can't think of them as people with souls. And there's a Joni Mitchell song about, you know, when she suggests that love is touching souls. Well, in the modern era, it's harder to think of such a romantic idea.

10:18Well, you push back on that in the book, because doesn't a character say, you know, there's sort of, you can explore a room, but there's rooms within rooms within rooms. Am I remembering right? I know, that's very good. Yes, I was relying on you to push back. I was just playing devil's advocate. No, I mean, in our story, both in the book and the film, I mean, we don't have any clear conclusion. And yes, I mean, in the book, there is a passage in which somebody suggests that actually, however brilliant Clara the robot is, however super intelligent, and however well-equipped she seems to be at observing and learning another human being and everything about them and the way they think and the way they feel.

11:07So, it's more like entering a house with many rooms, and when you enter each of those rooms, you'll find more rooms, and there'll be more rooms. And that maybe is what the human heart is like. And it will be very difficult for any machine or any system, you know, however able to ever get to the bottom of what a human heart is. There's something forever unfathomable about the human heart.

11:44But then Clara says, this machine says, well, yeah, I think maybe I'm up to it. You know, I'll try very hard. We'll see about that. We have to take a break, and when we come back, we'll talk about how much has changed in AI in the five years since Kazuo Ishiguro wrote Clara and the Sun. And how he thinks fiction can help us make sense of a changing world. Don't go away.

The rapid pace of AI

12:22Clara and the Sun was set in the future. It felt when I remember when I read it. Like, I remember very distinctly where I was. You know, it's a good book when you remember where you were when you were reading it. And I remember thinking, oh, this is, you know, this is really distant future. It doesn't feel that way so much now. I'm curious for you, has the pace of change of AI surprised you?

12:45Well, yes and no. I mean, it has kind of surprised me. Because when I started to write this book, people weren't really talking about AI. To some extent, I'm surprised people aren't talking about gene editing more, both in terms of the enormous benefits that's bringing in terms of health care and other things, you know, about food and control of disease. But also, you know, I think it does pose enormous threats and questions about how we organize our society.

13:16If it becomes possible to produce, you know, superior children, you know, people who are objectively superior, not just, you know, you're thinking that my son is nicer than somebody else's child or something. I mean, these are very challenging, challenging thoughts. What's happened in recent, in the last, you know, four or five years is that AI has come to dominate people's consciousness.

13:46I'm still not convinced that we're asking exactly the right questions about AI. I can understand that, you know, certain groups of people feel their particular livelihoods are threatened. So they suddenly start to make a lot of noise, you know, and that's as it should be. But I still think there are some big questions about AI that's not to the fore. Well, what are the questions we should be asking about AI?

14:20Yeah, well, I always feel uncomfortable when I get asked that because you've got to understand, Flora. I mean, the Nobel Prize is given to many very clever people, but I'm not one of them. Do people treat you, let me just as a side question, do people treat you like an oracle? Like, is that what, you know, because you... I was given a Nobel Prize for making up these kind of slightly strange stories. I'm watching this whole thing that we've been discussing about the rapid progress of science and technology in these fields.

14:54I'm watching from a seat way, way back in the auditorium. There are certain dangers to asking people like me questions like that because there's a danger that we start focusing on the wrong emphases about these things. Because, you know, I don't really know, you know, what's going on, you know. Having said that, it's also equally dangerous to just leave it to the experts because this is something that impacts people at every level, at every day level.

15:33And once again, I come back to our story, Clara and the son. I mean, it takes place in a domestic setting. It's about one particular family and what happens when you introduce a superintelligent AI robot into that life. You know, we can't just leave it to experts to pronounce and control the parameters of the debate. It's, you know, we have to all participate in this.

16:06You know, everyone has to say, look, we're worried about this or, you know, reassure me about this or how do we protect that?

16:16At the moment, there aren't many obvious platforms, it seems to me, where we can actually democratically, in the most general sense of the word, you know, have that debate. I mean, I think this comes back to what you said earlier about emotional truth versus scientific truth. And, you know, obviously we love scientific truth on Science Friday, but I also think emotional truth is important. You know, I'm going to go out on a limb there.

16:47I think that's important, too. And that is why I'm interested in your point of view on AI from wherever you are in the auditorium, because I think it might help us understand what seems to be a profound thing happening right now. Yes, I think emotional truth is very important. I mean, that's how I used to justify what I did, you know. I mean, if you sit on the stage with a bunch of Nobel Prize winning scientists, and then you have to go up and get your own Nobel Prize, and you start wondering, why am I here with these guys?

17:26So, do I have anything like what they have, you know, and you cling to this idea, oh, maybe I'm getting it for emotional truth, rather than this kind of stuff. But the trouble with emotional truth is that it's harder to debate in a disciplined way. I mean, in science, there seems to be something that has been developed over centuries, you know, the scientific discourse with experiment, peer review.

18:01There is a more or less agreed way of actually having your arguments, even if there isn't a clear-cut conclusion to them. With emotional truth, it's much messier. People are just saying, but that's my emotional truth, you know, well, too bad, I feel this. And so, we try to have a kind of a critical dimension in the cultural world, it seems to me. We try to say, look, that is just emotional manipulation, because it lacks some kind of emotional truth.

18:37It might make you laugh and cry, you know, it might be very transporting. But there's something about it that is not true.

18:47You know, we don't have a clear system like you have in the world of science. And so, it can be very dangerous. And someone who can wield emotional truth very powerfully can manipulate electorates and nations for their own ends. And so, it becomes a very real thing, and it becomes a very dangerous thing, when it's not regulated, or when it's not, when we don't have a kind of a context, as you do in science, a kind of a disciplined context to actually evaluate claims of emotional truth.

19:25That's such a fascinating idea, and I can't think of the antidote.

Hidden biases and political risks

19:34But, I mean, all right, but having said all that with massive qualification, you asked me, you know, what are the other questions about AI? I mean, what I worry about still, about AI, I worry about the prejudices and biases that will disappear within black boxes in AI. And we will feel compelled to come to certain judgments, some of those judgments that, it can range from people being denied the loan to people being sent to jail or even executed.

20:06And we won't quite know why we're depending on the conclusion reached according to things inside a black box. And I'm very aware that throughout history, we had things like slavery, which was supposed to be perfectly okay. And until very recently, and still in many parts of the world, I mean, a gay person is thought to be someone who should be put in jail, or worse.

20:41So I'm sure if you start to build black boxes now, it's going to be full of the equivalence, today's equivalence of those things. Except it's going to be much harder for us to go back and unpack those things, as we did with things like racism and homophobia, and indeed, things like slavery. But we know that human society changes, and by and large, I think, you know, I believe, progress is for the good.

21:19But I think that that's going to be very difficult once we leave all the decisions to things inside, deep inside a black box that we can't even analyze. And I'll tell you the other thing that, in my kind of cultural world, I would interest me. It interests me, and it also frightens me. And it touches on what we said before about emotional truth. If you get to a point where an AI can actually produce, say, a novel, or maybe a movie, that makes people laugh and cry, and really transports them.

21:59Quite aside from the question about whether that's taking work away from real human beings, when an AI program can do that, it seems to me it's in a very good position to be used politically. When you have an AI system that can really understand how best to emotionally manipulate people through their emotions, get them to do things, and really target particular demographics, and basically identify exactly the thing that will get people worked up, and compel them to do what you wish them to do.

22:47I think that that would be a very dangerous use of AI.

22:55You'll have to come back and talk to us for your next book. Will you?

23:01Yeah, but I have to warn you, my next book has no intellectual content whatsoever.

23:10It's coming out in March, and I've written it purely as an entertainment thing. Look, what I'm hearing is that it's another way of exploring emotional truth, and we're interested in all kinds of truth on Science Friday. Well, I'm really pleased to hear that you're interested in emotional truth on Science Friday. I have to say, I think it's really important that there are—I said earlier that we seem to lack platforms where we can debate these things so that we don't just leave things to the experts.

23:43But I think it's terrific that there are shows like yours and stations like yours where people can go. Everyone can go to actually—this is where we connect. I mean, there is a slightly Citadel-like feeling about elite science, that people feel we're not allowed in, and people talk in strange jargon, and we're not qualified even to step through the doors. So, you know, I think it's wonderful and really important what you're offering us now.

24:17Well, that makes my day. So, thank you very much. Kazuo Ishiguro is a novelist and Nobel laureate based in London, and the film adaptation of his book, Clara and the Sun, comes out in theaters today. Thank you so much for joining us today. This was really my pleasure. Thank you, Flora. It's been fantastic talking to you.

24:40This episode was produced by Kathleen Davis, and special thanks to our friends at NPR's studio in London. Thanks also to you for listening. I'm Flora Lichtman. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you.