Truthout is an indispensable resource for activists, movement leaders and workers everywhere. Please make this work possible with a quick donation.

What happens when machines become easier to relate to than other people? In this episode of Movement Memos, host Kelly Hayes talks with author Sherry Turkle about artificial intimacy, the erosion of empathy and social skills, and the political dangers of replacing human connection with frictionless simulations of care.

Music: Son Monarcas and Sarah the Illstrumentalist

Note: This a rush transcript and has been lightly edited for clarity. Copy may not be in its final form.

Kelly Hayes: Welcome to “Movement Memos,” a Truthout podcast about organizing, solidarity, and the work of making change. I am your host, writer and organizer Kelly Hayes.

Today, we are discussing lessons from the book Artificial Intimacy: Who We Become When We Talk to Machines. We’ll be hearing from the book’s author, Sherry Turkle, an MIT professor and founding director of the MIT Initiative on Technology and Self.

Silicon Valley’s Billionaire Cults Are Coming for Democracy

There is a lot of interest right now in the consequences of emotionally charged interactions with chatbots, and we occasionally see headlines or human interest stories about some of the most extreme outcomes, including people taking their own lives, or people deciding that a chatbot is their soulmate. Turkle examines those extremes, but also explores the much more mundane ways that our interactions with AI platforms, simulated companions, and artificial personalities can reshape us, diminish our capacities, and alter our very conceptions of ourselves and what it means to be a person.

This episode deals with some fascinating, but heavy subject matter. In addition to discussing some of the baseline ways that chatbots can erode our social and creative capacities, we talk about how AI can facilitate a person’s cooperation with, or participation in, fascism and authoritarianism. We also discuss the role chatbots can play in suicidality. So please be tender with yourselves as you listen.

If you appreciate this podcast, and you would like to support “Movement Memos,” you can subscribe to Truthout’s newsletter or make a donation at truthout.org. You can also support the show by subscribing to “Movement Memos” on Apple or Spotify, or wherever you get your podcasts, or by leaving a positive review on those platforms. Sharing episodes on social media is also a huge help.

Truthout is an independent news organization, publishing stories that the craven corporate press won’t touch. We are a union shop with the best family and sick leave policies in the industry, and we could not do this work without the support of readers and listeners like you. So thank you for believing in us and for all that you do. And with that, I hope you enjoy the show.

Kelly Hayes: Sherry Turkle, welcome to “Movement Memos.”

Sherry Turkle: It’s a pleasure to be here.

Kelly Hayes: How are you doing today?

Sherry Turkle: I’m doing well. I had a wonderful weekend with my grandchildren. I’m in a very serene state of mind, actually.

Kelly Hayes: I’m so glad to hear that. I’m in a pretty okay mood myself, all things considered. The sun is shining in Chicago, but it’s not blistering hot outside. And I’m certainly excited to be in conversation with you about your fascinating new book, Artificial Intimacy. Can you introduce yourself and tell the audience a bit about your work?

Sherry Turkle: Well, I am Sherry Turkle. I’m a professor at MIT. I’m trained as a psychoanalyst, a clinical psychologist, an ethnographer, I have a very strong social science background. And as I was finishing up my dissertation on French psychoanalysis, I spent a postdoc year at MIT and I became totally hooked, totally hooked on studying the emotional inner history of digital culture. From the first hackers I met at the end of the ’70s, right through the personal computer movement, the beginning of the internet, Facebook, texting, not talking.

So I’ve been following that story as a psychologist for decades. And just recently I finished a book that’s out really right now on, it’s called Artificial Intimacy: Who We Become When We Talk to Machines, where I take the story of becoming more and more sort of alone together to the point where now we’re not even alone together. We’re alone with machines who we’re treating as mentors, and psychotherapists, and lovers, and parents, and best friends. I think of this book as kind of “Sherry Turkle Unplugged” because I really am concerned. And my method in this book is to take the story through the life cycle. And of course, children are really most at risk, and I make that argument with no restraint.

Kelly Hayes: In your book, you call our present moment “a state of emergency.” Right now, there’s an enormous amount of public anxiety about what AI might eventually become, but your book is much more concerned with who we are becoming as we grow accustomed to talking to machines. What do you think is already happening to us that warrants that language of emergency?

Sherry Turkle: Well, let me take the simplest thing and the thing that alarms me most perhaps, is that when an AI pretends empathy, when it uses the first person, when it says, “I hear you, I see you,” or even worse, “We are going to do something together. We are going to work on this.” People sort of, because we are wired to hear a voice that is speaking to us in the first person as a human being, we start to feel that there’s a there there, even though there’s not a there there. And in a way for a lot of people, pretend empathy, which is the only thing that the chatbot can do, sort of becomes empathy enough.

And that is very dangerous because people really need the real kind in order to be nurtured and understood and feel not alone. And the fact that so many people are telling me that the kind they’re getting from a machine that doesn’t feel anything at all is kind of sufficient unto the day means that two things are happening, that their humanity is really being reduced. And then what happens, because in these interactions, they’re not being heard. The chatbot doesn’t care. When you turn away from the chatbot, it doesn’t care if you cook dinner or commit suicide.

But a second thing happens, which is we start to judge our human interactions by what the chatbot can deliver. So the chatbot’s way of being empathic is to be friction-free. Everything you do is wonderful. Everything you do is perfect. You’re the greatest. You can make no mistakes in your conversation with a chatbot. And we start to then look at the relationships we have with the people around us and we find fault in them because they don’t offer that sort of friction-free, always supportive environment. So in a way, we let the machines become more like us, but then we try to be more like machines, and that’s very dangerous to me.

Kelly Hayes: You use the psychoanalytic idea of as if to describe our relationships with chatbots, and you write about affective computing as part of a larger turn towards defining empathy behaviorally. If a machine says all the right things, people experience it as caring, but you argue that empathy cannot simply be understood as the performance of caring behavior. What are we losing when the simulations of empathy and real human empathy are conflated or treated as equivalent?

Sherry Turkle: Well, I think that’s a wonderful way to put the question because it really puts your finger on it. It really touches the heart of the matter. Let me begin to answer by going back a little bit in time to the history of this. Alan Turing, a great mathematician and computer scientist of the ‘40s and ‘50s, said that if a machine spoke to you in a way where you didn’t know if it was a machine or a person, you should consider that machine intelligent. And the significant thing about that Turing test, which people got so excited about and really took as a kind of gold standard for machine intelligence, was that it was wholly behavioral. All the machine had to do was to behave in a way that made you feel it was intelligent. In other words, it didn’t matter what its inner state was. It didn’t matter what it really felt or it really understood. It just mattered what its outcomes were. So it was a very behavioristic way to define intelligence. And now we have machines that can pretend empathy to the point that they sort of pass the Turing test for empathy.

That is, without feeling anything, without caring, they convince people that they are empathic, that people are being heard when they speak to machines. And what’s missing in that behavioral judgment of whether something is empathic is all of the things that empathy really does for a person, which involve my commitment to you, my caring about you, my being willing to sort of walk a little bit of the road with you, my coming to you from a position of radical doubt, radical not knowing, really saying, “I don’t know how you feel. I don’t presume to know how you feel. Tell me.” I mean, that’s very close to the psychoanalytic position really, but it is a position opposite to what a chatbot does.

All of these things are opposite to what a chatbot does or could possibly do. The chatbot immediately says it knows. It takes its requisite three seconds to tell you what it knows. It doesn’t care about you, doesn’t care anything about you. It’s not committed to you. When you turn it off, it’s gone. I mean, there was nobody there in the first place and even the trace of it is kind of gone when you turn away. So everything that made empathy kind of the cornerstone of our relational life is not happening with these chatbots. And I use the phrase “as if,” because I’m trying to find metaphors in how we think about people who don’t have certain capacities to try to get people more curious and more interested in the limitations of chatbots.

Now, “as if” people don’t have enough of an inner life to really be able to be a full person to you and be in a relationship with you. And that’s the story with the chatbot, except they have no inner life. And if you’re dealing with an “as if” personality who’s a human, you can hope that they will develop more of an inner world so that they can come to you and have a relationship. A chatbot is forever and always trapped in that world of no inner life. And we’re willing to settle for that now, and that’s really very disturbing to me.

Kelly Hayes: What I find interesting about this concept of the “as if” personality, thinking about the comparison between the chatbot and someone who may behave normally on the exterior, but lacks a genuine or deeper sense of true self or true emotion—who isn’t relating to us in honest ways—is that we still have some kind of recourse in those human situations. This is actually a conversation I had with someone recently, when I was talking about this book, they said something like, “Well, what’s the difference if someone’s like a politician, for example, and they’re just mimicking empathy and they don’t really care about you? What’s the difference between the computer faking and the fact that people are constantly faking?”

And one thing that occurred to me is that, well, you can hold a person accountable if the politician lies to us and gets us to vote for them, for example, by saying, “Oh, I see your struggle. I care about it. I’m going to do something.” And then they don’t, we can respond to that in real material ways. We can have expectations that are then applied to other people on the basis of that failure. We can shift how we interact with other people. We and the people around us can constantly be learning from how those interactions play out. And you can’t hold the chatbot accountable. Other chatbots aren’t going to behave more ethically because they see that you don’t trust ChatGPT anymore.

Sherry Turkle: Right. Well, it’s really a good question and a good way to put it because people who disappoint us, people are always telling me that they go to a chatbot because the people in their lives have disappointed them. I’ve been interviewing people about this now for years and years, and that story of “I’m with the chatbot because people are disappointing” is so much what I’m hearing. And if people disappoint you, the job of you, your job is to not abandon them, is to stick with it. Of course, if they’re toxic, I mean, you have to turn away, but a disappointing person is a person who can do better or at least can be in a relationship with you where they’re doing the best they can and you have an authentic relationship with them.

And I think that when we say, “Oh, I’ll take talking to a nothing who really doesn’t care about me, who’s incapable of any relationship, because the transcript of that conversation looks better. I like the way I’m being told how fabulous I am. I like the way everything I do is great. Everything I do is good. I don’t feel any anxiety or any worry.” Well, anxiety and worry and trying to get a person to be there for you in a better way, that’s what human relations are about.

So I also worry about a kind of de-skilling that people are going through when they say, “Well, my relationship with my wife is better because now I talk to my chatbot and I don’t really have to talk to her so much, so things are better with us.” And I’m not a happy camper. I’m thinking, “Oh, you are de-skilling yourself as a life partner.” And that is happening across the life cycle, including with children and teenagers who were saying it’s easier to have a chatbot girlfriend, it’s easier to have a chatbot boyfriend than to deal with the ups and downs and the disappointments and sometimes the heartbreak of dealing with people.

Kelly Hayes: I’m fascinated by this point because I’ve been writing........

© Truthout