Making Sense - Sam Harris - November 03, 2026


#501 — The Performance of Empathy

#500 — The Playing Animal

Episode Stats


Length

23 minutes

Words per minute

161.55

Word count

3,785

Sentence count

204

Harmful content

Hate speech

3

sentences flagged


Transcript

Transcript generated with Whisper (turbo).
Hate speech classifications generated with facebook/roberta-hate-speech-dynabench-r4-target .
00:00:00.000 you're listening to making sense with sam harris this is the free version of the podcast so you'll
00:00:06.420 only hear the first part of today's conversation if you want the full episode and every episode
00:00:11.400 you can subscribe at sam harris.org there are no ads on this show it runs entirely on subscriber
00:00:18.020 support if you enjoy what we're doing here and find it valuable please consider subscribing today
00:00:22.980 i'm here with sherry turkle sherry thanks for joining me my pleasure it's great to finally
00:00:28.840 connect with you. I don't, am I right in thinking that we have not met anywhere, not at a TED
00:00:33.540 conference or? No, I think this is the first time. Okay. Well, great. Well, I've been an admirer of
00:00:39.280 your, you know, the perspective you've given on technology for many years now. You've written
00:00:44.000 many books. The latest is Artificial Intimacy, Who We Become When We Talk to Machines. And we'll
00:00:49.700 get to AI, which is your focus there. But I mean, you've been thinking about our relationship with
00:00:53.600 technology, you know, as instantiated by computers and smartphones and social media
00:00:58.920 for a long time. If you were going to compare the world of today to, say, 1975, right, when
00:01:07.180 most people had nothing more than, you know, a television set and a landline, how do you think
00:01:13.620 we're doing? I mean, how do you perceive the effects of information technology and all its
00:01:19.660 guises permeating and subsuming human culture to the degree that it has?
00:01:26.500 Well, I think I'm going to narrow that question down to my field of expertise, which is the
00:01:31.860 development of the person and how the person situates themselves in society in terms of
00:01:38.420 feeling a sense of agency, a sense of community, et cetera. And there, there's been a remarkable
00:01:44.640 deterioration in how we relate to each other, in our confidence in relating to each other.
00:01:51.640 We've developed a feeling of vulnerability so that we want to escape into our phones or
00:01:58.180 don't text, don't talk, just text me. Don't text me. I'll talk to an avatar. I mean,
00:02:05.240 first we talked to each other. In 1976, we talked to each other and then we talked to each other
00:02:11.500 through machines. And now we're kind of dropping the middleman, which is the story of my new book,
00:02:17.120 and we just are willing to just talk to the machines themselves. So that's really a downgrading
00:02:22.900 in our kind of capacity and interest to deal with the rough and tumble and vulnerability
00:02:30.180 of dealing with other people. And similarly, we've retreated from our communities.
00:02:36.840 When I first started studying personal computers, for example, the people who were building them were interested in using them as kind of neighborhood hubs and ways of, their slogan was, use the virtual to improve the quality of life in the real.
00:02:54.220 That was the slogan of the early personal computer people.
00:02:58.060 Whereas Mark Zuckerberg now talking about meta basically says, people would like to have 15 friends, but that's unrealistic.
00:03:05.880 They're too busy.
00:03:06.980 The demands of their jobs or two jobs are too great.
00:03:10.480 So they can maybe have two or three friends, and he'll give them the extra 12 or 13 friends
00:03:16.640 as avatars, as programs, as chatbots that'll be tuned to them.
00:03:22.280 So even there, you see a kind of willingness or a corporate interest in saying, well, we're
00:03:29.340 not really going to be members of our community.
00:03:31.400 We'll even drop the real people to just look at screens.
00:03:35.140 So I think it's, you know, when I began looking at this phenomenon, I was optimistic about using the virtual to improve the quality of life in the world.
00:03:45.380 Can you recapture some of that optimism?
00:03:47.220 And what was your first interaction with the internet like?
00:03:50.460 I mean, were you an early user of, like, were you on the well with lots of people?
00:03:53.840 I mean, the people who I was, I wasn't myself, I was on the well, but I wasn't an active user of the well.
00:03:59.600 But I studied the well and I hung out with those people.
00:04:02.660 And the idea really was to create community online that would translate into more and better community offline.
00:04:12.240 And I remember visiting the early personal computer owners in San Francisco and a group called Community Memory that Lee Felsenstein started, a great pioneer of the personal computer movement.
00:04:24.780 And that was their mantra, improve the virtual to improve the quality of life in the real.
00:04:32.220 Make the virtual to improve the quality of life in the real.
00:04:36.560 And, you know, now there's kind of a retreat from the real and a sense that if your best
00:04:44.520 friends are chatbots, if you turn away from the kind of biological people in your life
00:04:50.200 to the easier to deal with virtual people, well, that's kind of understandable because
00:04:57.420 real life is very hard.
00:04:59.460 So actually in this book, I really come down very hard on that because I think that's a disaster for the individual, for communities, for children as they grow up.
00:05:09.600 It's a disaster for politics.
00:05:11.680 And I say, don't abandon the real now.
00:05:15.380 You know, reality really needs you.
00:05:17.520 This is not a good time to abandon the real.
00:05:19.960 And I feel that very strongly.
00:05:21.980 So what is the evidence that we have lost many of these capacities?
00:05:26.360 Were you saying, why was I optimistic?
00:05:29.440 Well, no, I mean, I just, I love to listen to, I mean, I often listen to people speaking
00:05:35.800 from the period where most of us were just getting online.
00:05:41.820 I mean, like there are people, I'm fans of various people who, you know, read their books
00:05:46.080 or let's just, there's an endless number of, you know, YouTube conversations and lectures
00:05:50.100 You can hear from people in, let's say, you know, 1994, who were, you know, early adopters online.
00:05:57.760 I'm thinking of someone like, you know, Terrence McKenna, who, you know, for all of his quirks as an intellectual, he was quite brilliant and quite prescient with respect to, you know, many of the cultural changes that would soon follow once so much of culture went into cyberspace.
00:06:13.000 But there was a kind of optimism there that really seems quaint and naive, but it's easy to be sentimental for that moment where you where we thought this has to be intrinsically good to have access to more and more of the world's information and to connect people with less and less friction.
00:06:34.680 Like on some level, what could be wrong with this?
00:06:37.560 I think the notion, the problem there is the word intrinsic. With technology, nothing really is intrinsic because right away, the social, the political, the economic environment in which the technology is being shaped jumps in.
00:06:54.500 So, for example, Facebook. I remember when I first heard about it, you know, grandmas look in pictures of babies. My classmates from all over the world had a kindergarten reunion. I had a fifth grade reunion. I mean, I did. I had a reunion of the people I went to fifth grade with. Unthinkable.
00:07:15.180 And so there was every reason to feel, my goodness, I guess the word intrinsic would come up.
00:07:20.660 You know, how can it be bad to connect the world with each other?
00:07:24.720 Well, here's how it can be bad.
00:07:28.540 American corporate culture actually thought of some ways to make it bad.
00:07:32.540 How can we make this bad?
00:07:33.900 Right.
00:07:34.820 Banner ads and algorithms.
00:07:37.980 Powerful minds went to work immediately.
00:07:40.080 And you can make it bad if you make attention the most important thing about it, keeping people staring at their screens.
00:07:48.900 And then you say, well, how do you do that? 0.89
00:07:51.200 You make people angry and then keep them with their own kind, silo them with their own kind. 1.00
00:07:58.980 That turns out to be like a really powerful way to keep people at their screens. 0.99
00:08:04.200 Now, if you do that, you get them at their screens, and you can give them ads, and you can scrape their data, and you can make money.
00:08:10.960 On the other hand, it's really terrible for the individual who's angry all the time, sort of activated, anxious, not going out, not meeting people, not getting nervous about the rough and tumble of the world as it is.
00:08:27.560 And it's terrible for them politically. It's terrible for the community, because what is a worse thing for democracy than to have everybody angry and only comfortable with their own kind and not wanting to talk across difference?
00:08:43.540 so something that sounded like a great a great thing it didn't take very long for it to be
00:08:51.320 a toxic thing and that's even before we get into the things that i'm kind of a specialist in which
00:08:59.340 is that you lose the capacity for solitude that people really need the capacity for solitude
00:09:06.060 in order to have mutuality you know you need to feel i know who i am i'm comfortable with who i
00:09:13.160 am. And then I can, in that comfort with who I am, I can say, well, who are you? And I'd like to
00:09:20.920 know you. And here's who I am. And let's have a relationship. If you don't know who you are,
00:09:25.820 essentially, you approach another person and say, tell me who I am. And that's a disaster
00:09:32.520 in terms of mutual friendship. So we gave up that. We gave up our capacity for boredom,
00:09:38.920 which is so important developmentally. People need to be able to let their minds wander
00:09:44.580 and have a moment of kind of inner peace in order for really their neurons, basically for the brain
00:09:53.480 to develop an autobiographical sense of self. The brain needs that downtime that comes with
00:09:59.220 boredom. Well, people get anxious if they're not being constantly stimulated. So we're really
00:10:05.580 taking away very important and taking away solitude and taking away the capacity for
00:10:11.000 boredom, where we're taking away very important signal developmental possibilities that we
00:10:18.400 need.
00:10:19.020 So all of these things are not good.
00:10:21.740 What's been most informative of your view of all that we're losing here?
00:10:27.640 Is it just, is it your interaction with students at MIT over the years, or is it the research
00:10:33.260 literature such as it is?
00:10:34.580 No. Well, the way I work, I mean, it turns out for the past 40 years, I've been interviewing people about their relationships with technology. Now, that's a very particular kind of methodology, but it's mine. And so really, I have, you know, 40 years of interviews about what it's like to sit in front of a screen, what people do, what their fantasies are, what they're doing, how they feel about themselves. I have that.
00:11:02.420 And then I have, you know, a wide reading in the research literature and other with other methodologies. But certainly on the things that I care about most, which is the loss of a capacity for empathy. That is my work supported by a very substantial, much more quantitative set of research studies.
00:11:27.060 You know, actually, there's a great story about empathy. I was discovering a loss in empathy when the cell phones became very big, 2010, 11, 12, 13, 14. And because people would say to me, you know, I'd rather text than talk. I'd rather be on my phone than talk to another person. And I'd say, why? And they'd say, conversation, I'll tell you what's wrong with conversation. It takes place in real time and you can't control what you're going to say.
00:11:57.060 In other words, the job of doing the empathic work was starting to feel scary, like too much to do.
00:12:06.320 And at that time, I read a study by a brilliant, a very quantitative study by a very brilliant psychologist named Sarah Conroth.
00:12:14.840 And I went to see her and she said, yeah, from the time of the cell phones on, you see this drop in all the ways that we know to measure empathy.
00:12:23.760 And I said to her, and she says, I'm very depressed about it.
00:12:27.060 And I said, well, what are you going to do? I'm depressed too. And she said, I'm going to build
00:12:32.200 empathy apps for the iPhone. In other words, I'm going to build technology for the iPhone
00:12:37.900 that I hope will give people practice and empathy, you know, tell stories and how do they feel and
00:12:44.240 how do they... And I thought to myself, this is a very bad time we're going into where technology
00:12:51.680 helps to create a problem. And then we look to technology to help us solve the problem.
00:12:58.460 And at the time I said to her, and I've said it many times in my writing, I say, but we are the
00:13:04.260 empathy app. People are the empathy app. And she laughed. I mean, she knew what I was saying,
00:13:10.340 but basically she did develop these empathy apps for the iPhone. And that's kind of, we're in the
00:13:16.000 same place today. Facebook and social media and chatbots have done so much to increase the crisis
00:13:23.720 and loneliness. And then all the programs say, well, here, we're going to give you chatbots to
00:13:28.500 fix the problem and loneliness. You're lonely? That's great. We'll throw a chatbot your way.
00:13:33.540 Well, let's linger on empathy for the moment because, I mean, first of all, you can mean two
00:13:39.360 things by the term. I mean, there's an emotional strand of empathy and then there's a cognitive
00:13:44.940 one. So cognitively being able to understand other people's experience is one thing. And that's,
00:13:51.640 you know, it's obviously a necessary piece of our cognition. I mean, this is, you know,
00:13:56.140 otherwise known as theory of mind. And some people have obvious deficits in that. But then there's
00:14:01.780 the emotional permeability that we have to one or another degree where we care about, you know,
00:14:08.560 when someone's sad and we see the external effects of their sadness, there's an emotional
00:14:16.220 contagion component, which their sadness begins to feel like our own. And I mean, some of that's
00:14:23.180 good. Some of it is the stuff of which real intimacy is made and also just the appreciation
00:14:29.120 of vicarious lives through art and film and et cetera. So we enjoy emotional contagion,
00:14:36.820 But there are also clearly pathologies of it, right? And social media is now this kind of, you know, cultural charnel house of, in some senses, too much emotional empathy, certainly too much emotional contagion.
00:14:50.740 I mean, there's a kind of—I mean, at the center of every moral panic is just the kind of unreasoning zeal provoked by certain images being taken, you know, as seriously as they can, right?
00:15:06.360 I mean, who knows even what the images actually depict in the real world, right?
00:15:09.580 Well, but I have a narrower definition—I think it's dangerous, personally, to have that inclusive a view of what empathy is.
00:15:19.080 Yes, how do you think about empathy?
00:15:19.880 I think empathy is that because two humans have lived the arc of a human life, know what it is to feel pain, know what it is to fear illness, know what it is to feel joy or have desire, or even the experience of being little and everybody around that you're dependent on is big.
00:15:44.320 and then as you get bigger, you start to have more agency, but you still remember and have
00:15:50.880 some of the anxieties of all those years when you were little, to look forward to things,
00:15:56.860 to lose someone, to be rejected. I mean, all of the things in a human life, I think when two
00:16:03.280 human beings face each other, they connect on these, I don't think of it as contagion,
00:16:11.520 I think of it as signaling to tune in to these moments in the arc of life and how they make us feel so that when you're hearing another person's story, you really can get into their head.
00:16:25.860 Now, I think that that is only good. And what you see online, which is whipping people up into a kind of, you know, almost pornographic, I mean, you know, every art form, every trick is being used to incite violence, incite hatred, incite tribal feelings.
00:16:51.760 I don't particularly think of that as the dance of empathy. I think there are other ways that have more to do almost with propaganda and mass delusion that have to do with that.
00:17:04.000 The thing that we're losing that I think social media and chatbots are undermining is really much more that person-to-person empathy where you really care about other people and you feel in a kind of community with them.
00:17:22.420 And my point in artificial intimacy is that chatbots can't do that, not because they're bad or because they're not well enough developed. They never will because they haven't lived the arc of a human life.
00:17:36.040 So they can have information about all these parts of a human life, but they haven't lived it. They don't feel it. They don't have the equipment to empathize.
00:17:50.980 So they perform empathy. And really what a lot of my new work is about and what the book is about is what happens to people when the performance of empathy becomes empathy enough. And that really is the problem that I say we face today, that for so many people, the performance of empathy by a program becomes empathy enough for them.
00:18:16.840 Well, that's an interesting distinction. I'm not sure I believe that distinction is so hard and fast, even in the case of human beings. But let's jump into the topic of AI.
00:18:26.680 Well, it's interesting to talk about because I think it is the question.
00:18:30.140 Yeah.
00:18:30.380 Because if you say, well, there's cognitive empathy and there's that personal stuff. Well, then it's easy to say a program has cognitive empathy. It's easy to kind of grant the program a kind of understanding.
00:18:45.180 because of its literature searches and, you know, it's sort of reading all the literature.
00:18:51.160 But I don't think, I think that's a, you know, empathy can have, the way empathy has two parts
00:18:58.200 is the way in which people, when they love each other, have two parts. There's the cognitive part,
00:19:03.640 my God, I admire you and you're beautiful. You know, there's the cognitive things,
00:19:08.360 you know, we want the same, we have the same values. And then there's the whole body experience,
00:19:13.960 falling in love, wanting to be close to the person, wanting to hold their hand,
00:19:18.160 wanting to touch them, make love to them. I mean, and you could say, well, let's just
00:19:22.300 separate out the cognitive part from like all the other stuff. But it turns out that love
00:19:27.960 doesn't work that way. And I don't think empathy works that way either.
00:19:31.680 But don't you think we could be humbled by a complete understanding of the way empathy and
00:19:37.840 sympathy and compassion and love work and how they're contingent upon certain surface features
00:19:43.940 that can be emulated by machines, ultimately.
00:19:47.720 So let's just imagine we get out of the uncanny valley
00:19:50.900 and we have humanoid robots that just,
00:19:52.840 I mean, we have Westworld, essentially, right?
00:19:54.860 I'm sure you saw Westworld, right?
00:19:56.600 Many times, many times.
00:19:58.720 Yeah, yeah, yeah.
00:20:00.280 Some of those seasons were great.
00:20:03.320 And so just imagine, I mean,
00:20:05.600 if you compare a perfectly tuned humanoid robot
00:20:08.700 to an imperfect person,
00:20:11.540 And I think many people will have the experience in the presence of a robot of feeling greater empathy, you know, in both directions, better understood, even more love than when they're in the presence of a person who, for whatever reason, you know, take a, you know, a person with Parkinson's who's got like kind of facial masking where they just, they're not expressing any emotion on their face, right?
00:20:35.680 because they're incapable of it or um you know somebody who's just uh you know for whatever
00:20:40.940 reason just not very charismatic and we have we have the charisma function turned up to 11 on our
00:20:47.000 perfect bot well you know you can feel things that aren't there in other words you can feel
00:20:52.800 that someone is performing empathy and is doing this this is no fantastic performance taking place
00:21:00.560 but in fact the robot doesn't the robot or the chatbot or the doesn't care when you turn away
00:21:07.560 let's just take chatbots for a minute because we're with the chat but it could we could make
00:21:11.060 it seem to care right like all of this is just what we but the question is whether or not it
00:21:15.340 matters whether it's performing seeming or actually care so i want to i want to just back
00:21:21.340 up to what you said because i think that it's a it's very provocative my position is is that the
00:21:27.960 performance of empathy is never empathy enough. Because if you have a person, I think it's
00:21:34.060 important to take out of this equation people with dementia, people with Alzheimer's, people with
00:21:40.040 Parkinson's. I was thinking of Parkinson's, but the kinds of Parkinson's that bring you dementia.
00:21:46.760 Because it's sort of, I'm willing to take out of the equation. I think it's important to take out
00:21:52.540 of the equation. People who really have had the parts of the brain that allow them to do some of
00:21:59.200 the empathy work removed. And take a person whose life history and whose family values and whose
00:22:08.940 way of being brought up has not made empathy and listening and conversational skills and reading
00:22:14.800 faces a very big part of their life. I mean, we know a lot of people like that. And then you say,
00:22:19.920 well let's compare that person which i thought you were sort of trying to do
00:22:24.740 with a really fabulous chatbot robot you know the most finely tuned it's the one that eric
00:22:31.240 schmidt is perfect you know it's yeah it's his favorite one for having for having chats
00:22:35.940 and my feeling is is that the most imperfect person has lived a life and knows pain and
00:22:44.840 knows fear and knows what it is to, you know, to be little and maybe fears death and doesn't
00:22:51.740 want to be sick and was worried about their children. And no, no chatbot feels that.
00:22:58.940 Members can hear the full conversation by subscribing at SamHarris.org.
00:23:03.160 Subscribers get a private RSS feed you can use with your favorite podcast player.
00:23:07.700 I think of myself as the front lash of the backlash. I think I make a very strong argument
00:23:13.040 that this is not good for us.
00:23:14.560 It's not good for our kids.
00:23:16.400 That they have trouble enough.
00:23:17.980 They're frightened of relationships.
00:23:19.620 They're not good at talking.
00:23:22.600 They haven't been given practice.
00:23:24.520 They're on their screens too much.