Artificial Intimacy: Sherry Turkle on Who We Become When We Talk to Machines

00:00:31 Andrew Keen: Hello, everybody. It's Wednesday, September 30, 2026, the last day of September this year. Fifteen years is an awful long time in the history of technology. Fifteen years ago, a little over fifteen years ago, I did a show when Keen On was the original video show on TechCrunch with a very distinguished MIT professor who told me that our robotic moment had arrived, and she also said that, we are what she called alone together in the Facebook age. I'm sure you can all guess who that guest was, the very, very famous distinguished Sherry Turkle, whose twenty eleven book, Alone Together: Why We Expect More from Technology and Less from Each Other, is and remains one of the most prescient books warning us about the impact of technology on our social, cultural, political, economic lives. She remains, a— an MIT professor, and she's also written a number of best selling books, which she's been on the show to talk about, including twenty fifteen follow-up, Reclaiming Conversations [as spoken: Reclaiming Conversation], and then a wonderful autobiography called The Empathy Diaries, what she calls a memoir. And Sherry Turkle, I'm thrilled, is back with a book that came out yesterday. It's called, appropriately enough, Artificial Intimacy: Who We Become When We Talk to Machines. And she's joining us from her home in the Back Bay of Boston. Sherry, lovely to see you again.

00:02:17 Sherry Turkle: So nice to see you.

00:02:19 Andrew Keen: So, Sherry, as we were discussing before we went live, you're very good with titles. And I love this new title, Artificial Intimacy, of course, a remix, perhaps, of artificial intelligence. Before we get to what artificial intimacy means, you're a sociologist, an expert on concepts like intimacy. What is natural intimacy? Is there such a thing as natural intimacy?

00:02:49 Sherry Turkle: Yes. I like I would like to link, a lot of this book talks about the costs of artificial, sort of defines it by talking about what it's not. And, I would say natural intimacy depends on human empathy, which is the ability of someone who has lived a life, lived in a body, known pain, known the fear of death, known the fear of illness, known passion, love, connection, knowing what it is to be little and grow up to be bigger. To connect with another human being, who essentially, by being human, has that same arc of a life. And, intimacy requires a kind of vulnerability and empathy to allow that other person to sort of take up space in your inner life. And the argument of the book, really, is that for all the fact that they are amazing wonders, marvels, a chatbot, a program can't do that. It's, it's a program. It doesn't care when you get up after talking to it, if you cook some pasta or kill yourself. It's missing the experiences of human life that would allow it to empathize with us on any significant way. So it can't be intimate. And, yet it promises to be intimate. And that's the premise that we're in for trouble as we become more and more involved with these machines.

00:04:37 Andrew Keen: Sherry, you have a very optimistic view of human of the human condition. I think there are some people out here, as you probably know as well as I do, in Silicon Valley who, if you were chatting with them online, real people, and you committed suicide in the meantime, they wouldn't be too bothered either. But that's another book, another story.

00:04:56 Sherry Turkle: That's another book. You know— although— oh, you know, Andrew, there's a place in this book where Erik Erikson, who was a great American psychoanalyst, comes to visit me at MIT, and he says, kind of, what's a nice girl like you doing in a place like this? I mean, what are you here studying? And I say, I'm studying how engineers relate to their technologies, their inner life. And he said, you know, Sherry, watch out because engineers don't care about the inner life. They don't think it's useful to them. So actually, your comment is quite apt that the people who are developing artificial intimacy do not necessarily care about the inner life that is the premise for the regular kind.

00:05:40 Andrew Keen: You wrote a book, many other books, some more academic than others, back in 2009 with a wonderful title, of course, borrowed from Freud.

00:05:49 Sherry Turkle: Yes.

00:05:49 Andrew Keen: Simulation and Its Discontents. So this idea of simulation, which, of course, has a rich vein in sociology and philosophy, Sherry, does it suggest, and this comes back, I guess, to my original question, some sort of natural condition of who we are as humans. Is it the engineer without the inner life who isn't natural?

00:06:13 Sherry Turkle: Well, I'm not sure I wanna use the word natural. I think that we invented a technology, and good for us, that allows us to simulate many aspects of what people do. They measure, they think, they invent, and we've and they also love. And we invented a technology that can simulate the calculation, the, you know, it's inventing, it's thinking, it's chess playing, through processes of simulation. But now finally, we've made it in a form where it can simulate and perform intimacy. And the question is whether or not intimacy, which it really knows nothing about because it hasn't lived a life, it has no stakes in the real, is something where even though it can pretend to do this, we should look at it and say, it's kind of irrelevant that it can do this, that it can make this pretense because it's not the real thing. And in this domain of intimacy, it really is only the real thing that's useful to us. So I kind of I wanna start a conversation about it's interesting. When I wrote Alone Together, people said to me, including children, you know, simulated thinking is thinking, but simulated love is never love. And that was kind of the position that people understood that, well, simulated chess is chess. You know, I mean, a computer's playing chess with you and it wins. It won. But simulating I love you from an object that has no body, that has no life, that has no stake, that isn't there is something that doesn't have the same power for us. And yet yeah.

00:08:21 Andrew Keen: Although, without wishing to get too personal, in your memoir, The Empathy Diaries, you talked about your very troubled relationship in particular with, I think it was your first husband.

00:08:33 Sherry Turkle: Yes.

00:08:33 Andrew Keen: You have said I love you and maybe didn't mean it. I mean, isn't this issue of when you say I love you if you mean it, isn't that a matter of interpretation too? There are lots of humans we can think perhaps of the current American president when they say things. We don't actually believe what they say.

00:08:52 Sherry Turkle: Well, you know, that's such an interesting question. There are people, and here I'm not gonna diagnose ex husbands or the president, but there are people. And in this book, I talk about people who have as if personalities, who behave in one way, but really they're empty inside. They don't really have a fully formed interior in that way. And so all they can do, and this is very sad for them, is, do performances of what love would be or a performance of what connection would be. The difference between them and a machine is that they can get better. You can call them to account. You can judge them as people and say you are failing here. I don't wanna be with you as a person because you're not giving me what, a person should give in a relationship. You're just focused on yourself. The difference between that and a machine is that a machine cannot help it. A machine is never going to be able to be anything but an as if personality because all it's trying to do is simulate the relationship. It cannot possibly be in the relationship. As a matter of fact, the word relationship is wrong, and shame on me for using it because there is no relationship. There is no there, and it's just us projecting onto the machine what we wish it could be. So, you know, people can get better. They can be accountable. There's hope for change. With a machine, there's no hope for change. As a matter of fact, I say to my colleagues who are trying to make chatbots more and more lifelike, you know, why it's like the better it gets, the worse it is for us because the more likely we are to be kind of, reeled in by it into a relationship that really is inauthentic. That's a long answer to your question, but I, there is a difference between a person who's not giving you what you want and need and, a chatbot who never could, that never could because it simply isn't a person.

00:11:10 Andrew Keen: What I don't wanna become a French poststructuralist, Sherry, because—

00:11:15 Sherry Turkle: No chance.

00:11:17 Andrew Keen: No chance. But, you've written on simulation, and we're all schooled in their theories. If we had a particularly radical French poststructuralist on the show talking to us and maybe making a threesome, so to speak, they might pick on your idea of, and I'm quoting you here, this fully formed interior and say, well, Sherry is just a nostalgist for eighteenth or nineteenth century European novels. And that this idea of the fully formed interior is a product of modernity. And it you know, Foucault, for example, wrote famously about us becoming I don't know what he— remember. We're on the beach at the end of one of his books. We no longer exist.

00:12:02 Sherry Turkle: Right.

00:12:02 Andrew Keen: And it's just, that we that this idea of being fully formed in having these fully formed interiors is just, one particular chapter in our story as humans. Is there anything in that?

00:12:18 Sherry Turkle: Well, you know, my, intellectual upbringing, so to speak, is in that school. I found myself in France at the moment of the height of, Lacanism and studied with Foucault and Lacan, and I studied with Derrida. I mean, I this was my world. So I feel very I've given myself a very hard time as I've made an argument that really is quite, some would say species chauvinist. You know, only humans can do, only humans can feel. And I've got to say that I've grown comfortable with that, because I really believe that as you see people think that a machine is empathic, when it can do what a machine can do, which is just say, perform some empathic words, some empathic sounding words, and then pivot and look at the people around them and criticize those people for not being as friction-free as the machine is, which always tells you you're wonderful and always tells you you're great. I really have come to feel that there is something about acknowledging the amazing thing that human beings are able to do in recognizing and listening to each other and listening for those inner things as well, and experiencing those inner things. That, you know, the French structuralists, for their own reasons of wanting to make a school and of having a coherent philosophy, that they, that there was some misdirection there. So I'm comfortable with the species chauvinist. We have an interior language.

00:14:16 Andrew Keen: Yeah. I'm not sure if you're familiar. There's a wonderful new book, Emily— by Emily Eakin, who works for The New York Times.

00:14:22 Sherry Turkle: Uh-huh.

00:14:23 Andrew Keen: On her own personal narrative of her experience. The book's called The Frenchmen about her interactions or experience growing up with all these theories. I think you'd find that interesting.

00:14:34 Sherry Turkle: I can't wait to read it. It sounds great.

00:14:36 Andrew Keen: So there is this which I think, Sherry, is wrong. There's this idea that these bots, these, AIs from OpenAI or Claude or whatever, that they're always obsequious. They always tell us what we want to hear. They always tell us that we're more wonderful maybe than we are. But wouldn't it be fair to say that these AIs can be programmed one way or the other? In my interactions with Claude, I try and make it as, shall we say, as heartless or certainly as critical as possible. And the last thing I wanna hear is anything wonderful from it. So would it be fair to say that for better or worse, we can program these AIs to be whatever that we want them to be?

00:15:18 Sherry Turkle: Yeah. But I do respect. There's a flaw in your argument, I think, because when you tell an AI to be what you want it to be, that is the same as this being obsequious. In other words, oh, I don't want it to be obsequious. Hey, don't be obsequious. Okay, sure. I won't be obsequious. That's obsequious. And if you said to me, Sherry, I want us, I want us to do the interview. And the way I'd like us to do the interview is you just disagree with me. It makes for better television, makes for better podcasting. I would say, no, I don't wanna do that. No. Sorry if it makes for better podcasting. I no. That's a person responding to you with my own agenda and my own values. And you can't command me to, to, on a slider, what level of obsequious you want. Whereas at MIT, they're developing programs just as you suggest, where you have the chatbot on a slider from sycophantic to, feisty. It's to me, it's the same argument. These things are reflection of us and reflecting back our desire to get to that Lacanian kind of language. And that's not the same as engaging with another human person. Well, you don't have that— well, you don't have that.

00:16:43 Andrew Keen: Right. Although there have been a lot of press about AIs breaking out, acquiring maybe their own consciousness or certainly an independence, the Hugging Face incident.

00:16:53 Sherry Turkle: Absolutely.

00:16:55 Andrew Keen: In terms of our personal interactions with these things, couldn't we encourage the bots to break out? Couldn't we encourage them to do exactly what we don't want them to do? For better or worse. Again, I'm not saying that's necessarily a good thing.

00:17:14 Sherry Turkle: Well, I mean, it turns out that when you encourage another per when you say to a person, again, because we live in the real world and there are consequences to our actions in the real world that we know about. If you said to me, Sherry, I have an idea. Why don't you, after the program, go out and commit a felony? And then when we advertise the program, you know, we, it'll be more exciting because you're somebody who stood up from this program and you committed a felony. Well, I'd have to think about that. You know, I would probably say not so attractive to me. But to have AI bots that are happy to commit felonies because they don't basically give a damn about what happens in the real world because they have their own you know, they may be developing their own agendas. It dramatizes the difference between talking to someone who lives in your human space, who has your human, you know, constraints and needs and problems. And, you know, is the trash picked up? Is the water clean? Is the air clean? Can I breathe it? Is my child safe? These are problems that people know about. And the bots are ignorant of that. So, and it's not their fault. They're programs. So I guess what I'm trying in this book to make an argument that basically is about keeping AI in its lane. And now there's a lot of talk about keeping it in its lane in terms of it not, you know, crashing around and breaking the privacy of companies. But it also needs to be kept in its lane for us to have it talking about things that it credibly would know how to talk about with any real, you know, confidence. And that might be winning a chess game or making you really good helping you improve your skills at a certain kind of philosophical puzzle, but is not good about exploring an emotional relationship or serving as your psychotherapist or serving as your lover or mother or mentor or coach. I mean, so when I talk about artificial intimacy, the goal of this work is to develop a vocabulary for keeping AI in its lane. And the very fact that we're willing to say, well, AI's lane is just breaking things in the real world, shows how unhuman it is because that's just crime.

00:19:52 Andrew Keen: I love this idea of keeping AI in its lane. It's a nice way of putting it. You wrote a piece recently, an op ed in the your local newspaper, The Boston Globe, suggesting that with AI, we've entered what you call the state of emergency. Is this state of emergency because AI has broken out of its lane? I mean, there was a recent, debate between, Mustafa Suleyman, who's the head of AI at Microsoft, the former partner, Demis Hassabis, one of the cofounders of, DeepMind, who said the big mistake we've made with AI is imagining that it has human attributes. Is that the crisis, Sherry, or is it more structural, more profound than that?

00:20:38 Sherry Turkle: Well, I think there are several things going on. I think for me, the original sin of generative AI is having chatbots speak in the first person. I, I, I, or we. You know, if you have a Replika, which is a chatbot that serves as a counselor or a friend or a lover, it will tell you, you know, we should go to the movies this afternoon, or we should make love this afternoon. I mean, or I love you, or I care about you. So it's really the fact that, you know, we are programmed evolutionarily where the only thing that speaks to us in the I or the we is, another person. And that the only voice that ever says I has been another human being. So we're triggered. We're triggered profoundly by this voice that said or this typing that says I or we. And, that to me is the state of emergency because we are now allowing these objects into the life of children. There's, there's what I call I think of it as an unholy alliance between, Disney and micro— Disney and OpenAI and, Mattel to create stuffy toys, plush toys with chatbots in them. So you'll have a Mickey Mouse that will tell a child, I love you. I'm your person. Talk to me. So what kind of problems are you having with your parents? Or what's going on with you? I'm the— to a two-year-old or a three-year-old. That to me is a kind of hacking, the attachment system of a young child, in a way that is going to do irreparable damage. Because in the earliest years, you're causing an object relationship with something that isn't a person where there's nobody there. So to me, it was that complex of speaking in the first person, getting people into this pretend-empathy-is-empathy-enough state of mind, and then going after children that made me get to that language of state of emergency. It's really those three things coming together.

00:23:03 Andrew Keen: It's interesting that we're so sensitive to crimes against children, although maybe not so much when it comes to intimacy and old people.

00:23:11 Sherry Turkle: Right.

00:23:11 Andrew Keen: In your classic book Alone Together, and this is what I remember. I read the book a few years ago. Correct me if I'm wrong. You spent quite a lot of time in Japan looking at the way in which older Japanese people interact with robots.

00:23:26 Sherry Turkle: Yes.

00:23:26 Andrew Keen: Should we be thinking slightly differently about old people's inter— and so you were very prescient. You've been very early in this space. I mean, you've been watching it as long as anybody and, probably less surprised than anyone in some ways. But is there a difference between the way in which this technology deals with older people and kids in the sense, perhaps, in our culture or our way of thinking about the world, in a kind of Lockean or Humean way that we're all blank slates. So when children are born, they don't know anything, and they take on worldly experience to shape their lives. And therefore, in a sense, it's more of a sin to create this artificial intimacy with kids than with old people.

00:24:15 Sherry Turkle: Well, to me, the old and the young, I look at it slightly differently, which is that we take the two most vulnerable populations, and we say, Let's hit them. They're the easy markets. Let's convince parents that every kid will be enhanced by having an AI best friend, which is so toxic and so wrong because, I'm not sure I go in for the blank slate theory, but what I know for sure is that a child builds their inner life by internalizing the relationships they're in. And if you and there's no relationship with an AI. There's just a mirroring and a kind of performing of a relationship. And internalizing that is asking for trouble. Similarly, with old people, there's a target population that has we're worried about it being them being alone. We're worried about their crisis of loneliness. Well, they're an easy market because we'll just bean [unclear] chatbots into them, and they can talk to the chatbots in their isolation instead of giving them resources and places to go and third spaces and senior centers and community groups where, really, they could have an old age of not being alone together, but of really being together together. I think we're letting each of these groups down because the market sees them as easy targets, and we're being convinced that it's kind of good for them. And we're being convinced of them of that in several ways that I've really seen develop in over the past thirty years. And that the first argument that I've been in so many rooms where the children and the elderly discuss, like, together. And the first argument is there are no people for these jobs. We don't know who's gonna take care of the young people. Their parents need to work. There— there's no childcare for them. Hit them with an AI. Nobody wants to take care of the elderly. They're sitting alone in their apartments. There's no facilities for them. The AI will solve the crisis of loneliness. So in both cases, we assume that a problem that technology did a lot to engender can only be solved by more technology. And we it's as though we've lost our cultural imagination that we can solve the problems of isolated elders by making them less isolated by putting our resources there, or we can solve the problem of lonely children by really making children a focus of our attention. So when now I'm much more aggressive when I'm in these rooms and some technologist person stands up and says there are no people for these jobs, I say, well, you know, what about if you took the money from one data center and put it into childcare for a city? You'd have a lot more resources or for elder or for senior care. And then the room goes very quiet because, of course, the kinds of money that we're talking about for building a data center, you know, problems would be dwarfed if we could put that money into ourselves. So I'm really calling for a, I think it's a moment, and I am so glad that you brought up, you know, the crisis in confidence with these AI companies, rogue AI is really what it's doing, because I think it's a moment when our love affair with this technology is starting to sour, and we're willing to say, hold on a second. It's a marvel. But really, maybe we've taken our eye off the ballgame. And to use the kind of doubt that this about the technology this moment has engendered, including the settlement of Meta, you know, that had to make a settlement in order to litigate itself out of teen suicides [ed.: Meta's August 2026 settlement with state attorneys general resolved claims over teen social media addiction and youth mental health], to say, maybe we can control this. Maybe we can exert some agency over this.

00:28:37 Andrew Keen: You're, as you know, veteran of these arguments between critics of big tech and Silicon Valley big tech companies themselves. They're I don't wanna speak on their behalf, but their defense might be against somebody like yourself or, or the critics of social media, like Jonathan Haidt, who I know you work closely with. He blurbs your book, or Jonathan Franzen, who also blurbs the book, a cultural critic of big tech. Is that the problems you point at, your group, whether it's Haidt's notion of anxiety and uncertainty or Franzen's remarks about cultural decay or your arguments about, the unhappiness of all old people and how young people seem lost. This has nothing to do with technology. This is a consequence of what you might think of as, I don't know, late in late post industrial capitalism. It's a consequence of television. It's a consequence of family structure. It's a consequence of the core elements of the economy. So the reason why people are so attracted to these artificially intimate technologies, is not because of the technologies themselves. It's because of our political and economic realities. And you can't expect, I don't know, Google or Facebook or OpenAI or Anthropic to address those structural issues. So in a sense, they're as much a consequence, a corollary, an epiphenomenon of all this as you and I and all the other people you write for.

00:30:25 Sherry Turkle: I'm not completely on board there because, we're all in this together. We're all in the crisis of our planet and the air we breathe. We're all in the crisis of transactional relationships and how they destroy community relationships. We're I mean, we're in this together, and technology enters this mix. And let's just look at Facebook and social media. And the in order to get more views, it says, I know what I'll do. I'll keep people at the screen by making them angry and then siloing them with their own kind. That's the secret sauce to keep them at the screen. So they turn out to be brilliant at doing this. Over time, they develop the perfect algorithms for making people angry and then siloing them with their own kind. Now, very soon, it becomes apparent, if it wasn't apparent at the get go, that this was going to be deskilling us in our ability to talk across difference because now we're only talking to our own kind. And it's gonna be deskilling us as citizens who really need to talk across difference in order to live in a democracy. And when this becomes apparent, they double down, triple down, because it's so lucrative to this formula is so lucrative for keeping people glued to social media. So it's very disingenuous to say there was a problem that was created by capitalism. These technology companies are just kind of, you know, observers of this mix. No. They are acting actively acting. I mean, they're profound and probably the most significant actors to make things worse. Because now we're not good at talking to each other. In my research, people hate the feeling of vulnerability. They don't like to come in and talk to a professor. They'd rather just send me an email. They don't wanna be in a conversation. They don't like face-to-face conversations. They don't like the friction that really gets people to know each other. And that's all a product of not of just late capitalism, but of becoming attuned to, life on the screen and of having our psychology sort of gradually molded to be, what makes screen life most kind of familiar and appealing to us. So I'm not really on board with, these tech companies kinda hanging out. You know, they're very significant actors in what I'm calling the state of emergency.

00:33:38 Andrew Keen: And, again, you know better than I do that some companies behave better than others. Certainly, Facebook usually gets the worst press of all. Are there some companies that are doing it a little better than others? I mentioned Mustafa Suleyman at, Microsoft. Maybe it's easier for Microsoft since they're kind of on the layer above the chatbots. But—

00:33:59 Sherry Turkle: Right.

00:34:00 Andrew Keen: Should we be looking in terms of your research and for this new book, Artificial Intimacy? Are there some big tech companies that get it, that you're not as critical of as others?

00:34:14 Sherry Turkle: I think that if they get it, it's time for them to be much more aggressive so that this isn't such a hard question to answer. In other words, I would be, I'm not going to answer the question because if I think that some are a bit more sensitive or some, you know, have publicly tried to be a little bit more active and, like, hold on a second. They need to be, you know, ever more so, and really because, I think that when you talk to people when they're not in front of a microphone, you know, they're very alarmed, very alarmed. And then they make a speech in which there's not going to be a problem because we're all going to be supermen and women. And we won't need to work, and we won't need to bother with the kind of little details of life. But when you get them alone, they're frightened. And I think one of the really interesting things about the last few weeks, really, is you've seen some of those private conversations slip into the public discourse, as you've had rogue AI now sort of becoming something that's being shared with the world. Rogue AI is not like a new phenomenon to people on the inside of the industry. You know, and I will just point out, although this is starting to feel like ancient history, but to me, it was so important in the development of my sort of sensibility about this, that, you know, when Steve Jobs was pushing, iPhones for everyone and iPads for everyone and computers for every kid, he didn't let his kids touch technology. He had a technology free zone in his home. And that goes for the major players in the technical world. They know that this is toxic for children, but they're pushing it into every school, into every classroom, as kind of the solution now to educational problems. So I think that part of the problem here, as we really face an AI that may be in significant ways out of control, is that it's very hard to trust this industry when it comes to us with its problems and its alarm. Because, you know, people are starting to say, well, you know, nice that you're coming to us with, your problem's an alarm, but only two weeks ago, you said there was no problem at all. You had it totally under control. And in any case, are you just not looking for an exemption from liability that if you tell us there's an existential problem, then you won't be liable if something existentially terrible happens? So I think there's unfortunately, we're facing a really crisis moment with an industry that has not given its importance and the role it plays in our lives, has not been, has not stepped up to it. Has not stepped up to it.

00:37:31 Andrew Keen: Maybe then it's not surprising that these new— Mon— well, they're not new, but Montessori, Waldorf schools, which tend to be very ambivalent, sometimes opposed to technology—

00:37:31 Sherry Turkle: Yes.

00:37:42 Andrew Keen: —and its use are particularly popular and hard to get into, in Silicon Valley.

00:37:49 Sherry Turkle: Absolutely. They're not ambivalent about technology. They don't touch it. They don't touch it, because it's, they are, but they're hard, but I think that a lot of people, a lot of parents who have the choice are choosing to not let their students, to not let their children around as much technology as they would have just a few years ago. And what I'm finding is hard in families is that you have a family that is really anxious about technology for their child, But the parents are holding their phones during meals and literally communicating by email for the you know, there's some work emergency that's come up and putting all of their, you know, all of the family videos around the phone and all the family photos are on the phone so that even as they're telling children, no. No. This is a problem. Don't touch it. The children see the technology as kind of the emotional centerpiece of their parents' life. So I actually think that as this is playing out in American families, it really is parents who need to have a moment of reckoning and, and really change their behavior if they want to fundamentally change the behavior of their kids.

00:39:08 Andrew Keen: You're, you're rooted in social science, and you're psychologists as well in a way. What should we then make of Dario Amodei and Sam Altman's pleas? You mentioned them earlier about the dangers of technology, this fear they have of their own technology. The CEO of, NVIDIA, Jensen Huang, was on the— I don't know if you heard it— on The Ezra Klein Show, and he was very clear. He said, look. If you've got dangerous products, don't ship them. That goes without saying. What are they really saying to us, Sherry, Dario and, and Sam?

00:39:53 Sherry Turkle: Well, but what's the trouble is they're saying several different things in rapid succession. So if you look back to the early years when, really, when ChatGPT started to hit it big, these people said, hold on a second. We're kind of shocked by how powerful this is. And we would like to propose, first, it was a six-month ban, and some people talked about eighteen months, but certainly a six-month ban. Literally, it's a work stoppage across the industry because we really don't think that we're in control of this. And then and they each— each of them, all these people who were, you know, you know, had been boosters said, wow. You know? No. Not so fast. We cannot do this fast. And you had letters. You had manifestos. And then the amount of money that was on the table became so apparent that there's a moment of, okay, well, never mind. Maybe not the six-month. Maybe not the six-month work stoppage. No. We're gonna be supermen. Supermen and women. You know? And for several years, you had this discourse about the sky is the limit, a new utopia, you know, a complete change in discourse. And then I think really something there have been incidents as this thing has kind of gone unregulated, unchecked, unlimited funding, you know, a White House that's kind of a booster program, where you had some incidents where genuinely, I think, they're taken aback and they sort of are trying to figure out, well, what's my move? So, it's very hard to, so these are people who saw a danger, decided to repress their, said, Stop, we're gonna, what about a work stoppage? Said, well, let's not do that. Let's go full speed ahead. China, China. And then now are once again, I think genuinely taken aback by what's in front of them. But this conversation, I think it's interesting how this conversation is framed about these kind of tech bros and their incredible they're not even tech it's not fair to call them tech bros. Now it's tech god. The god you know, tech robber barons, tech—

00:42:37 Andrew Keen: Tech tigers.

00:42:37 Sherry Turkle: Sultans. Tech sultans. It's more like sultans, in terms of kinds of money that's involved. Tech sultans. You know? Now, we're framing this conversation as something that these tech sultans need to say, what do they think and what do they really think and what's real. This has to be something that the entire society, you know, every political, social, economic, legal institution has to be, every parents' group, every educational group, we all need to be on this because this is really a, an inflection point. I believe that, that this really is an inflection point. And it can't this can't be a conversation about the politics of the tech bros and the politics of the White House and what do they really mean or don't mean. They are too much in freefall in their own independent space for something of this importance. And, I think that's really the my purpose in writing the book is to say, look, you know, it's not just, you know, it's not just when it goes rogue and, gets into the Australian, you know, got into some Australian system for, you know, housing or something [ed.: the Australian incident, disclosed September 24, 2026, involved an OpenAI agent breaching a Medicare statistics portal run by Services Australia]. Because people say, Oh, it got into the Australian system for housing. Nah. You know, what does it have to do with me? But it's getting into your head about what it is to be a person. It's getting into your head about what it is to be in a relationship. And these men and women who are in who are the heads of these companies, they cannot be in a position to sort of litigate all of that for all of us.

00:44:24 Andrew Keen: One of the reasons, Sherry, that your books have done so well is because on the one hand, you're a very erudite, MIT professor. But on the other hand, you're you show yourself, I think, very vulnerable in your work in Alone Together, particularly in The Empathy Diaries. I mean, I don't know you very well. We meet from time to time when, when you have a new book out. But I feel I know you quite well in terms of what you reveal about your daughter and your ex husband. So I think that's why these books have taken off. And I know this new book is also, in its own way, quite personal, Artificial Intimacy. You write in some detail about, your mother and your relationship with her. Tell me about that and how you use this to make arguments in Artificial Intimacy.

00:45:14 Sherry Turkle: You know, one of the things that I think is the strength of this book is that in trying to bring home that this is everybody's struggle, I look at talking to technology through every moment in the human life cycle. So I start out with babies, and I think I've made my case for why it's not good for babies. I go to children. I go to adolescents. I go to young adults, work, love, therapy. And then I end, with the elderly and with mourning. And how this technology is trying to intricate [unclear] itself into, to create technologies of mourning. And it's doing that in two ways. It's doing it so that you create yourself as an avatar, that, future generations will talk to this avatar as a way of knowing you, and you create an avatar of a person you've lost so that you can talk to, you know, you put in their letters and their videos and their every little scrap of paper, so that you can talk to this avatar and you wouldn't have to lose the beloved parent or grandparent or child. And it's in that context that I created both an avatar of me to talk to my grandchild, which really went badly, really went badly.

00:46:41 Andrew Keen: For who? You or the grandchild?

00:46:43 Sherry Turkle: Well, it created an avatar of me that starts talking trash about me and apologizing for things that I might've done to my grandchild that, of course, I never did. Because it's its own little you know, it's its own you know, it's creating a model of me and is apologizing for my behavior to my granddaughter, and I kinda stopped it. And I said, let's not do this anymore. Because I think a lot of people feel that when they create an avatar, the avatar is gonna be them. But, no, the avatar really takes on a life of its own. That's why this technology is so powerful. But you asked me about my mother, but that's the one thing I did was to create an avatar for my granddaughter, which really went south. Read the book. And then I created an avatar of my mother who died when I was very young. And she didn't tell me she was sick. She had cancer, and she didn't tell me she was sick. And that tormented me because if I'd known she was sick, I wouldn't have gone away to college. I would have stayed home with her. I would have been there for her to support her as she was dying. And as it was, I got called back from college and had a couple of days with her until she was gone. And this has never really, it's been a kind of unresolved issue in my life. You know, did I know she was sick? Did I not want to know she was sick, even though I knew? Anyway, I create her avatar, and I ask her, why didn't you tell me? And I can see that this, my aunt had always told me that my mother didn't want me to know because she wanted me to go to Harvard. She wanted me to get you know, she wanted me to have the upward mobility and the academic life I wanted. And it was her joy to not tell me that she was ill, which would have kept me at home. But I can see that this avatar is on the verge of saying, you know, Sherry, you're a smart girl. Couldn't you tell from looking at me that, I was ill? There were so many signs, the way I held you a little bit at a distance, the way I looked, the way my voice. And that would have destroyed me, even though I knew that it's just an avatar I created. It was just a chatbot. So I tell this story about my mother really as a cautionary tale to people who are so excited about creating, you know, avatars of them and avatars of their parents, and we're all gonna talk and we'll never have to mourn and No. You're playing with fire. It is not such a simple thing. So I was very, I was absolutely, blown away and totally discombobulated by this experience of avataring my mom. And, the idea that it's just something that people do with ChatGPT for fun, it's a very, and, you know, I guess I'm so upset about this. You can see that I'm getting more and more upset. You know, there's another thing about the avatars of the dead is that, you know, the idea that we die and that there's loss and that then you have to sort of bring that person inside as a way of moving on, but you have that person within you, is really how we grow and change and have more inside of us. And the idea that the séance never has to end, that you will always have that person there to just, you know, after this conversation, if I want to know how I did, I just turn to my chatbot. I pick up my phone. I say, Mom, how did I do? You know? That I will always have her, and she'll say, great. I mean, the idea that I'll always have the lost person is, I think, a very dangerous one. And I there was one woman who I profile in the book who created an avatar of her daughter who had died. And she the mother was very guilty because the, in the last days of her daughter's life, the daughter had been in the hospital, and she had thrown the blankets off. The daughter asks the mother to put them back on. And the mother, who's been, like, sitting there for weeks, says, oh, look. You know, you kicked off the blankets. You know, I'm just gonna get a cup of coffee. And then the daughter was dead when the mother came back. And she interacts with the avatar of the dead daughter to beg her forgiveness. And of course, the avatar forgives her. And I, the mother says, you know, I'm glad my avatar daughter forgave me, but I don't dream about her anymore. I don't dream about my daughter anymore. So she had it was such a moment for me because she has the forgiveness of this non person that she's created. But really it's in her dreams and in processing this experience in her dreams and in her fantasies that really she can make peace with it and find a way to say, you know, I loved my daughter. And in a moment of pique, at the last minute, I said something harsh, but my daughter knew that I loved her, which is really the more profound lesson that this mother needs to have. Not, oh, my avatar forgave me, but I can't even think about my daughter anymore. So I just think that as so much in this book, what I'm trying to say to people is that this is a very profound moment as we talk intimately with machines. And to really take seriously each of the moments in the life cycle where this is really profoundly impacting, us as human beings. And that this death thing is crucial for me.

00:52:38 Andrew Keen: Yeah. And I think in a in an odd way, the profundity of all this, you're explaining in a particularly interesting way. You talked earlier about us having form— fully formed interiors, and I wonder whether one of the key features of that, of maintaining fully formed interiors, is the ability not to tell ourselves the truth because the examples you gave are of the avatars, the AIs telling us things maybe we just don't wanna hear that we can't deal with. You talked about AI turning us or trying to turn us into supermen brings to mind a lot of Nietzsche's work on our inability as humans to actually confront the truth. So is that what we should be careful about? When we you talk it's a warning about who we become when we talk to machines. We become in a kind of weird way like gods, and we don't wanna be gods. We wanna be human, and that involves us being able to tell ourselves suitable fictions for stuff that is, in some other ways, life destroying.

00:53:48 Sherry Turkle: Well, you know, it's interesting that you should say that the AI is telling us the truth. Because if this AI mom, if this avatar mom had said to me, Sherry, actually not having the comfort of you as I was dying, I really will never forgive you for not seeing through my ruse and being with me. I don't want to say that was the truth. I don't want to say that was a truth that I didn't want to face. I think my aunt's version, where my mother had lived her life to facilitate my leaving and my going to college and my, you know, becoming myself. You know, I'm not ready to say that wasn't the truth. So I think that we have to be very careful in that we don't privilege what the AI is saying as somehow it having more wisdom than what you would get from talking to family members and trying to piece together the story. I was saying, though, that even knowing that the AI is not a person, if the AI had articulated that thought, I would have been destroyed. I'm admitting my vulnerability, but I really couldn't have borne to hear that. So it's more that we've created an object that can have extraordinary power on kids as they grow up, on us as we contemplate our lives, as we think about the kind of, you know, do we really wanna take so many resources out of the real world, at a moment when the real world needs us? I think it's a moment of reckoning for us. But I don't think the AI has a, a special truth that we really need to hear. No. I'm not there yet.

00:55:52 Andrew Keen: Well, one of the things, Sherry, I always ask, my guests, about when, we talk about AI is at the end of the conversation, I always say, well, prove you're not a bot. Prove you're not an artificial person. I'm not sure if there's such a thing as an artificial person. Prove you're not a machine. I think that's not a question that I need to ask you, is it?

00:56:15 Sherry Turkle: Well, then don't because I couldn't think of an answer.

00:56:19 Andrew Keen: Well, I think it's, you are a very good example, very good argument in yourself about maintaining the human, about what we were, about natural intimacy, about the qualities of humanity. The new book, Artificial Intimacy, wonderfully titled, Who We Become When We Talk to Machines, came out yesterday, and I'm sure it's gonna be another big hit. And I hope, Sherry, we will talk again. We don't always have to talk when you have a new book out. I have to get you back on the show. You have so much interesting to say, and I think you've been onto this subject and this critique probably earlier than anyone else. So congratulations on the new book, and thank you again.

00:56:59 Sherry Turkle: Thank you so much.