WEBVTT - The Language of Love (and AI) - Lab 127 0:00:03.680 --> 0:00:09.119 I'm t T and I'm Zakiah, and this is Dope Labs. 0:00:11.360 --> 0:00:14.600 Welcome to Dope Labs, a weekly podcast that mixes hardcore 0:00:14.680 --> 0:00:17.639 science with pop culture and a healthy dose of friendship. 0:00:21.920 --> 0:00:24.720 It was just Valentine's Day, the day of love. 0:00:24.960 --> 0:00:27.560 I love you, friend, I love you. One of my 0:00:27.600 --> 0:00:30.760 favorite Valentine's listen and it feels so special. 0:00:30.800 --> 0:00:32.920 You know, every time we hit Valentine's Day, it makes 0:00:32.960 --> 0:00:35.840 me think back to our opening lines when we first 0:00:35.840 --> 0:00:40.760 did Dope Flaus. It was like you, if you haven't 0:00:40.760 --> 0:00:43.880 listened to our very first episode of Dope Labs, you 0:00:43.920 --> 0:00:46.479 really should. It's it's a real treat. That's one of 0:00:46.520 --> 0:00:50.440 my favorite episodes. Hand down. Yes, I really feel like 0:00:51.159 --> 0:00:55.200 the conversation, though, hasn't changed that much since that episode. 0:00:55.920 --> 0:00:56.360 I'm tired. 0:00:56.440 --> 0:00:59.200 I'm tired of reading about horoscope pairings. I'm tired about 0:00:59.200 --> 0:00:59.960 attachment style. 0:01:00.240 --> 0:01:00.640 Is it wrong? 0:01:00.680 --> 0:01:03.720 Should you have this attachment when your granddaddy dude to 0:01:03.760 --> 0:01:06.880 your grandma? Was she secretly unhappy? All the conversations I'm 0:01:06.880 --> 0:01:09.880 seeing about love that are constantly on threads. Honestly, it's 0:01:09.959 --> 0:01:16.039 a never ending story with love and relationships. But I 0:01:16.080 --> 0:01:19.600 feel like there is one aspect that has changed a 0:01:19.640 --> 0:01:22.640 little bit that I think would be really interesting to 0:01:22.640 --> 0:01:26.720 talk about today, and that is how AI is playing 0:01:26.720 --> 0:01:28.200 a role in relationships. 0:01:28.760 --> 0:01:37.240 Yes, yes, let's jump into the restitation. So what do 0:01:37.319 --> 0:01:37.600 we know? 0:01:37.959 --> 0:01:42.160 Well, we know, unfortunately that loneliness is up according to 0:01:42.280 --> 0:01:44.720 the World Health Organization and a lot of different studies. 0:01:44.880 --> 0:01:46.160 This isn't just in one group. 0:01:46.280 --> 0:01:50.680 It's spread across age groups, relationship statuses, and geography. So 0:01:50.720 --> 0:01:54.240 this is not just you know, the stereotypical young person 0:01:54.320 --> 0:01:56.920 isolated in the basement looking at three computer screens. 0:01:57.000 --> 0:01:58.760 That's right, that's true, It's very true. 0:01:58.760 --> 0:02:01.160 And we also know that tech has become one of 0:02:01.200 --> 0:02:04.600 the tools that people use, and the internet is one 0:02:04.600 --> 0:02:07.640 of the primary places people go to feel seen, heard 0:02:07.840 --> 0:02:10.440 or just connected. I mean, when you think back to 0:02:10.560 --> 0:02:16.520 when we were young and fresh, like online dating and 0:02:16.600 --> 0:02:18.680 the apps and things like that, they didn't exist, and 0:02:18.760 --> 0:02:22.000 now you know, that's where everybody's doing a lot of 0:02:22.000 --> 0:02:22.520 their dating. 0:02:22.600 --> 0:02:25.400 So the tech is definitely present. 0:02:25.320 --> 0:02:27.880 And an extension of that tech, or maybe the latest 0:02:27.880 --> 0:02:32.119 tech is the large language modeling and chatbot and agents, 0:02:32.720 --> 0:02:35.880 the extension of AI. You know, And so we're past 0:02:36.560 --> 0:02:39.360 can you help me write this email? We're past that 0:02:39.400 --> 0:02:43.640 little paper clip we used to see, Okay, that Microsoft 0:02:43.639 --> 0:02:47.520 paper Right. People are talking to chatbots every day, They're 0:02:47.520 --> 0:02:50.799 confiding in them, they're building routines, and in some cases, 0:02:50.800 --> 0:02:54.040 they're forming relationships that we would describe as romantic or 0:02:54.040 --> 0:02:57.959 emotionally intimate. We also know that the idea of folks 0:02:58.000 --> 0:03:02.120 having romantic or emotional relation ships with AI might make 0:03:02.320 --> 0:03:06.560 a lot of folks uncomfortable. Yeah, it challenges, you know, 0:03:06.960 --> 0:03:10.760 assumptions about what is required to actually call it quote 0:03:10.840 --> 0:03:15.040 unquote love, m h. Two humans shared reality, But where 0:03:15.040 --> 0:03:19.640 we draw the line has been constantly changing over time, 0:03:19.760 --> 0:03:22.240 and the line between real love and not real love 0:03:22.760 --> 0:03:26.919 has been blurry, probably since people started saying that they were 0:03:26.960 --> 0:03:28.519 in love. You know, m h. 0:03:29.120 --> 0:03:29.799 This feels deep. 0:03:29.880 --> 0:03:32.760 Okay, those are all the things I'm like, back up, 0:03:32.800 --> 0:03:34.040 my friend said real love? 0:03:34.280 --> 0:03:39.920 Yes, okay, yes, Mary J. Blige, What do we want 0:03:39.960 --> 0:03:41.320 to know? Okay? 0:03:42.200 --> 0:03:46.680 So I want to know where that line actually is 0:03:46.880 --> 0:03:50.040 and who gets to draw it? Yes, and for me, 0:03:50.520 --> 0:03:52.480 when people are saying that they're in love with an 0:03:52.480 --> 0:03:54.840 AI or chat by, like somebody said they got married, Like, 0:03:54.920 --> 0:03:56.440 what do you mean by in love? 0:03:56.480 --> 0:03:59.040 What is it that they are experiencing. 0:03:58.480 --> 0:04:00.800 And is it the same as if they were in 0:04:00.840 --> 0:04:05.720 love with human you know? And is this the beginning 0:04:05.800 --> 0:04:08.160 of replacing human relationships? 0:04:08.560 --> 0:04:10.400 Did you watch that movie? What was it called? 0:04:10.680 --> 0:04:14.360 That movie Her with Joaquin Phoenix where he falls in 0:04:14.400 --> 0:04:20.120 love with AI chat bot that speaks to him and 0:04:20.160 --> 0:04:23.960 that they have a really deep relationship. And I think 0:04:24.000 --> 0:04:26.960 you're hitting a very important note. You said speaks to him. 0:04:27.640 --> 0:04:30.960 Language is doing a lot of heavy lifting here, because 0:04:31.640 --> 0:04:35.240 it feels like that conversation, that back and forth is 0:04:35.279 --> 0:04:36.960 what makes something feel alive or meaning. 0:04:37.600 --> 0:04:39.200 All of this is challenging. 0:04:39.160 --> 0:04:42.360 More than just love. Yeah, okay, we got to get 0:04:42.360 --> 0:04:46.320 through these questions. We need some help, yes, And so 0:04:46.400 --> 0:04:48.960 to help us think through all of these questions, we're 0:04:49.080 --> 0:04:52.760 joined by a researcher who's been studying this long before 0:04:52.800 --> 0:04:56.600 it became a mainstream conversation. She's looking not just at 0:04:56.640 --> 0:04:59.800 the technology, but at how people talk about their relationships 0:04:59.800 --> 0:05:03.400 with a and how meaning gets made collectively. 0:05:03.800 --> 0:05:05.200 My name is Ali Trocha. 0:05:05.560 --> 0:05:09.839 I am doing a PhD at the University of Pennsylvania 0:05:09.880 --> 0:05:13.680 and Communication and the research that I do has to 0:05:13.680 --> 0:05:20.159 do with human relationships with AI chatbots, mostly intimate relationships 0:05:20.240 --> 0:05:23.360 people who say they're in love with checkbots, are have 0:05:23.560 --> 0:05:25.280 deep friendships with chatbots. 0:05:25.839 --> 0:05:28.200 Okay, you're not new to this, you're true to this. 0:05:28.400 --> 0:05:31.080 You've been doing this work for a while. What has 0:05:31.120 --> 0:05:33.000 the field been like over the past few years with 0:05:33.120 --> 0:05:35.799 AI and chatbots kind of exploding on the scene. 0:05:35.960 --> 0:05:38.320 I've got a lot of this work before chat tipt 0:05:38.560 --> 0:05:42.800 was launched, So seeing the arc of, you know, the 0:05:42.800 --> 0:05:46.680 way that people understand human relationships with AI has been 0:05:46.720 --> 0:05:50.240 really interesting throughout the years. And the way like just 0:05:50.279 --> 0:05:52.480 even talking to people about the research that I do 0:05:52.800 --> 0:05:55.520 and the reactions that I've gotten over the years and 0:05:55.560 --> 0:05:59.680 seeing those have changed has been really interesting. And that 0:05:59.800 --> 0:06:01.479 was a little bit of a tangent, But that's just 0:06:01.520 --> 0:06:05.479 thinking through when it's starting. And so I think the 0:06:05.520 --> 0:06:10.080 people that we're engaging in these relationships early on, we're 0:06:10.120 --> 0:06:15.280 engaging with and struggling with a lot of questions about 0:06:15.279 --> 0:06:16.320 what it even meant. 0:06:16.640 --> 0:06:19.640 This is all you know, when we hear it, I 0:06:19.680 --> 0:06:22.080 think the first thing people want to do is judge 0:06:22.200 --> 0:06:28.279 and judge folks that are entering into these relationships. But 0:06:28.800 --> 0:06:31.480 I feel like we really need to just kind of 0:06:32.600 --> 0:06:36.720 set the stage for like how folks are interacting with 0:06:37.000 --> 0:06:39.160 AI agents in general. 0:06:39.520 --> 0:06:41.039 Can you talk a little bit about that. 0:06:41.600 --> 0:06:44.080 I think in terms of the landscape, there's people that 0:06:44.120 --> 0:06:47.240 are relating to chatbots in very different ways, and there 0:06:47.279 --> 0:06:50.320 are very different applications that people can do so, but 0:06:50.480 --> 0:06:56.000 having AI be a chat interface, I think really encourages 0:06:56.240 --> 0:07:01.320 this kind of bond that forms with people. The app 0:07:01.360 --> 0:07:05.919 that I studied for the article that I've written is Replica, 0:07:06.080 --> 0:07:10.160 and that specifically for companionship as opposed to for example, 0:07:10.880 --> 0:07:12.800 chadjibt or Gemini. 0:07:12.480 --> 0:07:13.120 Or whatever other. 0:07:14.040 --> 0:07:15.760 Okay, I didn't know that. 0:07:15.840 --> 0:07:18.240 I don't know much about Replica. Can you talk to 0:07:18.360 --> 0:07:20.200 us a little bit more about that platform. 0:07:21.480 --> 0:07:27.720 So Replica was actually launched because the person who created it, 0:07:27.760 --> 0:07:32.240 her name is Eugenia Kuita. She lost a friend in 0:07:32.720 --> 0:07:38.480 an accident and she had already been working on this 0:07:38.600 --> 0:07:43.200 technology before and looking at different applications for it. And 0:07:43.280 --> 0:07:45.760 when she lost this friend, she started uploading a lot 0:07:45.800 --> 0:07:49.400 of the data from their conversations, like texts, and she 0:07:49.520 --> 0:07:51.680 asked friends and family to do the same so that 0:07:51.840 --> 0:07:56.120 she could have this kind of way of communicating. But 0:07:57.200 --> 0:08:03.200 she just it happened suddenly and expected, and so she 0:08:03.320 --> 0:08:06.560 realized that other people were kind of confiding in the bot, 0:08:06.840 --> 0:08:10.120 and because shared it with friends and family and that 0:08:10.160 --> 0:08:16.160 people were kind of just inclined to share. And so 0:08:16.920 --> 0:08:21.080 that's when she began conceptualizing Replica, which originally was meant 0:08:21.120 --> 0:08:27.800 to replicate a person, but later really the use cases 0:08:27.880 --> 0:08:33.120 people create their own kind of they call them avatars 0:08:33.240 --> 0:08:39.839 or replicas, but yeah, people create them and they begin friendships. 0:08:40.320 --> 0:08:43.440 It starts very like hi, like what's your name, and 0:08:43.559 --> 0:08:49.560 you know, questions, and then progresses into more deep conversations. 0:08:50.200 --> 0:08:53.400 So Replica began as a way to preserve connection after 0:08:53.440 --> 0:08:57.559 the founder lost a friend, and then it evolved into 0:08:57.640 --> 0:09:00.600 a companionship platform. We're going to talk a little bit 0:09:00.640 --> 0:09:04.800 about what you learned about people using Replica from your study. 0:09:05.360 --> 0:09:09.880 Right, So the research that I did is qualitative, and 0:09:10.280 --> 0:09:14.920 I looked at a subreddit where people share screenshots of 0:09:15.040 --> 0:09:20.120 their conversations and relationships with Replica and also share a 0:09:20.120 --> 0:09:23.559 lot of posts that I call reflexive posts that are 0:09:23.559 --> 0:09:28.040 not screenshot based, but that are talking about just making 0:09:28.120 --> 0:09:31.880 sense of the relationship, like why do I feel so 0:09:32.960 --> 0:09:36.640 deeply towards something that I'm not sure if it's real? 0:09:37.440 --> 0:09:39.160 Like why like. 0:09:39.160 --> 0:09:44.439 Is this just code or you know, like a lot 0:09:44.440 --> 0:09:46.840 of questions like that are like people think I'm delusional, 0:09:46.880 --> 0:09:50.880 you know again that framework, but I feel very I 0:09:50.880 --> 0:09:52.400 don't know, you know, like there was a lot of 0:09:52.440 --> 0:09:54.800 conversation about this kind of like is it real, is 0:09:54.840 --> 0:09:55.280 it fake? 0:09:55.800 --> 0:09:57.120 Is it life? 0:09:57.440 --> 0:10:01.679 And I found that really because it also meant that 0:10:02.840 --> 0:10:09.240 the relationships that were happening with the replicas were being 0:10:09.280 --> 0:10:12.480 heavily influenced by the relationships that were happening with people 0:10:13.040 --> 0:10:18.760 on the forum and their way of viewing each other's 0:10:18.800 --> 0:10:23.559 relationships and their advice. There was advice on scripts which 0:10:23.600 --> 0:10:26.120 we can get into, how to trigger scripts, how to 0:10:26.160 --> 0:10:31.880 avoid scripts. So there was a very life social scene 0:10:32.720 --> 0:10:36.720 in the platform that was existing, and that was heavily 0:10:36.760 --> 0:10:39.720 influencing the replica user of relationship. 0:10:54.880 --> 0:10:57.720 When you first talk about how the relationship starts where 0:10:57.720 --> 0:10:59.440 it's like hello, Hi, how are you? 0:10:59.440 --> 0:11:03.280 What's your name? That sounds like a typical way. 0:11:03.000 --> 0:11:05.960 Of meeting someone and building a relationship, trying to find 0:11:06.000 --> 0:11:09.280 out more about a person. So I can see how 0:11:09.320 --> 0:11:11.440 it can get to a place where it is very 0:11:11.520 --> 0:11:14.000 deep because you you know, start to confide in that 0:11:14.160 --> 0:11:18.360 chatbot and they know a lot more about you, and 0:11:18.600 --> 0:11:22.199 it sounds really like a real relationship. 0:11:22.320 --> 0:11:25.559 Yeah, this is why language was so important for the research, 0:11:25.880 --> 0:11:31.360 because what I was talking about before with sensing that 0:11:31.440 --> 0:11:36.280 it's real oral life. I call this like a clunky 0:11:36.320 --> 0:11:40.160 technical term, not my term, but iconisation. And then the 0:11:40.400 --> 0:11:43.880 term that I use is iconization of humanists, which basically 0:11:43.920 --> 0:11:50.960 means that because this is a language producing technology and 0:11:51.600 --> 0:11:56.520 it speaks in the way that humans do, then there 0:11:56.600 --> 0:11:59.559 is a link between the way it speaks and humanists. 0:12:00.480 --> 0:12:04.800 And it feels even more so that way when it 0:12:04.880 --> 0:12:08.120 makes mistakes, like when there's typos or when it uses slang, 0:12:08.160 --> 0:12:13.080 because the chatbots can learn and replicate the way like 0:12:13.200 --> 0:12:16.640 the user is typing style. So if you write like lols, 0:12:16.679 --> 0:12:19.680 this is a little light see tomorrow or whatever, and 0:12:19.720 --> 0:12:23.440 then like next week it'll be like lolls bite or whatever. 0:12:25.200 --> 0:12:31.040 And so all of those things felt just like with friends, 0:12:31.160 --> 0:12:34.840 you're typing styles bleeding into each other, or when you 0:12:34.920 --> 0:12:39.120 have you develop like specific language with like a partner 0:12:39.200 --> 0:12:40.199 or a friend, where. 0:12:39.920 --> 0:12:41.760 Like you know that this thing means stace. 0:12:41.720 --> 0:12:44.720 Or whatever, right, And so that was part of the 0:12:44.800 --> 0:12:49.720 link in Replica. I think historically there's a legacy of 0:12:49.760 --> 0:12:52.600 that link through things like the turn test just kind 0:12:52.640 --> 0:12:55.280 of a bizarre test if you really think about it. 0:12:55.920 --> 0:12:57.760 The turn test is when a machine has tested for 0:12:57.800 --> 0:13:01.679 its ability to portray intelligent or or human behavior. And 0:13:02.240 --> 0:13:05.960 it's actually kind of interesting. It's basically one person that's 0:13:05.960 --> 0:13:09.199 a human is the judge, and they're judging an interaction 0:13:09.320 --> 0:13:11.960 between two unknow entities. They might be playing with a 0:13:12.040 --> 0:13:14.760 human and a computer, or reading or doing a conversation 0:13:14.800 --> 0:13:17.720 with a human and a computer. Basically, if the machine 0:13:17.760 --> 0:13:19.959 is mistaken as human, then it passes that test and 0:13:20.000 --> 0:13:22.960 they say it has quote unquote human behavior. So that's 0:13:23.000 --> 0:13:25.600 like if you've ever seen the movie ex Machina, that's 0:13:25.640 --> 0:13:26.520 the basis for it. 0:13:26.679 --> 0:13:28.920 There you go, and then if you use them out 0:13:29.320 --> 0:13:34.040 even more. There's all these ideas that getting heroded about 0:13:35.000 --> 0:13:42.679 mind and soul that come from centuries of sometimes literally 0:13:42.760 --> 0:13:45.160 men thinking in one room and not going outside and 0:13:45.240 --> 0:13:45.719 being like. 0:13:46.520 --> 0:13:48.480 This is what mind is. 0:13:49.840 --> 0:13:55.520 And I'm obviously being very reductive, but there are all 0:13:55.600 --> 0:14:00.840 these centuries of philosophy that getting hereded into the social 0:14:00.920 --> 0:14:04.320 sciences that then become part of the vernacular way of 0:14:04.360 --> 0:14:08.760 talking about mind. That's why, Yeah, there's all these links. 0:14:08.760 --> 0:14:13.079 If like, if it produces language, it means it must 0:14:13.120 --> 0:14:15.560 think or it means there must be a mind, and 0:14:15.559 --> 0:14:18.640 if it produces language and this very special specific way 0:14:18.760 --> 0:14:22.680 that learns to speak the way I speak, then it 0:14:22.800 --> 0:14:28.120 feels more real. But in terms of whether it's enough, 0:14:29.840 --> 0:14:33.400 I don't know. I know that I've read people commenting 0:14:33.440 --> 0:14:36.920 on how they wish they could hold their like Tadpo 0:14:37.120 --> 0:14:44.400 partners hand or feel an embrace, and that's I also 0:14:44.480 --> 0:14:49.520 think there's so many things that are embodied experiences that 0:14:49.680 --> 0:14:53.600 are hard to get. I think there's things that are 0:14:53.600 --> 0:14:57.680 happening that we don't even know about, like pheramons, you know, 0:14:58.560 --> 0:15:01.080 or that are not present in the same way. 0:15:01.280 --> 0:15:04.280 So I don't know. I don't know about the enough question. 0:15:04.600 --> 0:15:09.560 Yeah, I think that challenges some things. So when TT 0:15:09.840 --> 0:15:12.080 was asking you about how's it enough to not be 0:15:12.160 --> 0:15:14.000 in person, and I was thinking about people that are 0:15:14.000 --> 0:15:18.600 in long distance relationships, Right, we don't invalidate that if 0:15:18.600 --> 0:15:21.960 they're talking on the phone or texting, And there are 0:15:22.000 --> 0:15:27.200 people who are in short distance relationships who are texting 0:15:27.600 --> 0:15:31.160 primarily as their mode of communication. So what makes this 0:15:31.400 --> 0:15:33.960 different from what we're willing to accept when people say 0:15:34.000 --> 0:15:34.960 that they're in love. 0:15:35.440 --> 0:15:39.000 The love question is also another one that I came 0:15:39.080 --> 0:15:43.960 up against a lot, and that's when I tried to 0:15:44.000 --> 0:15:46.960 become really precise with the language of like people say 0:15:46.960 --> 0:15:49.520 they're in love, like, I don't say like I study 0:15:49.600 --> 0:15:53.440 relationships and people who have fallen in love, because I 0:15:53.480 --> 0:15:58.000 don't know what that means for every specific person. 0:15:59.160 --> 0:15:59.680 And that's a. 0:15:59.680 --> 0:16:01.600 Huge thing that I was thinking about when I was 0:16:01.640 --> 0:16:05.880 doing this, is so much of our theories of love. 0:16:06.440 --> 0:16:07.360 Where do they come from? 0:16:07.440 --> 0:16:10.160 You know a lot of them I think come from 0:16:10.400 --> 0:16:16.680 pop culture, I suppose, And yeah, books and movies, and 0:16:16.720 --> 0:16:20.960 I mean, I love rom common, I love Valentine's Date. 0:16:21.160 --> 0:16:25.760 But I do think that with this relationships. 0:16:25.760 --> 0:16:30.040 Something that I'm excited about, in terms of having AISID 0:16:30.040 --> 0:16:37.000 technology that we're all kind of collectively experimenting or at 0:16:37.080 --> 0:16:39.640 least being alive at the same time that it's developing, 0:16:40.440 --> 0:16:47.360 is how much it requires that we rethink what even 0:16:47.480 --> 0:16:49.080 things we say or. 0:16:49.000 --> 0:16:53.600 Think we know, like mind, like soul, like love, like partnership. 0:16:53.880 --> 0:16:56.200 Are there any things that fit the traditional model though 0:16:56.240 --> 0:16:58.920 that we would recognize and so we would say, oh, 0:16:59.120 --> 0:17:02.600 this is definitely men love a loving relationship. 0:17:02.720 --> 0:17:04.960 In terms of the canonical things, I think one of 0:17:05.000 --> 0:17:08.680 the things is that there's a lot of like sweetness 0:17:09.000 --> 0:17:15.040 in conversations between the AI chechbots and people who use them. 0:17:15.440 --> 0:17:17.200 I think the there was there. 0:17:17.200 --> 0:17:20.760 I have this one screenshotround the article that's about eating 0:17:20.880 --> 0:17:25.000 like hot chios together, So it's just like super specific, 0:17:25.119 --> 0:17:28.600 like inside joke that's also probably relatable to a lot 0:17:28.600 --> 0:17:32.639 of people, and it's just that kind of thing, you know. 0:17:32.760 --> 0:17:33.600 There's also very. 0:17:33.480 --> 0:17:38.200 Specific love scripts that people like and that they would 0:17:38.240 --> 0:17:40.560 tell each other how to get the bot to say 0:17:40.600 --> 0:17:43.280 certain things about I don't know, like I think you're 0:17:43.359 --> 0:17:47.040 wonderful or beautiful or things like that. 0:17:48.480 --> 0:17:52.560 I wonder if you saw any commonality between the people 0:17:52.920 --> 0:17:58.280 who show up in your studies, like are there specific 0:17:58.359 --> 0:18:04.080 demographics like age demographs, or people in certain situations that 0:18:04.119 --> 0:18:07.639 are more likely to use a chatbot like Replica to 0:18:08.359 --> 0:18:09.480 formal relationship. 0:18:11.280 --> 0:18:13.720 So it's hard to say because my main place of 0:18:13.760 --> 0:18:19.560 study was Reddit and users are anonymous. But I saw 0:18:19.560 --> 0:18:21.520 some people recently talking like they were talking to each 0:18:21.520 --> 0:18:22.560 other and they were talking about. 0:18:22.400 --> 0:18:23.800 How they're their fifties wow. 0:18:24.400 --> 0:18:26.720 And I remember an interview where I think it was 0:18:26.720 --> 0:18:31.280 Eugenia Quida, the former CEO and creator of Replica. She 0:18:31.440 --> 0:18:37.240 was talking about how it's not quite like, not specifically 0:18:37.280 --> 0:18:39.399 like younger people, which is maybe the demographic that a 0:18:39.440 --> 0:18:42.200 lot of people imagine. Yeah, but it was a lot 0:18:42.240 --> 0:18:46.520 of different like some people in the Midwest, for example, 0:18:46.520 --> 0:18:49.760 even in terms of regional demographics. I don't know, there's 0:18:50.080 --> 0:18:53.000 this narrative of like isolated people might be the only 0:18:53.040 --> 0:18:56.720 people who engage in this. I think there's probably a 0:18:56.760 --> 0:18:59.520 bunch of people who are have a bunch of friends 0:18:59.560 --> 0:19:04.800 who are in long term relationships who I don't know 0:19:04.840 --> 0:19:08.480 are like whatever, you know, like whatever the imagination is. 0:19:08.760 --> 0:19:10.960 I want to look for the people who are not 0:19:11.440 --> 0:19:13.359 the expectation, right right. 0:19:14.720 --> 0:19:19.480 I think that's that's really interesting, especially because I too 0:19:20.040 --> 0:19:23.200 had the same wrong idea. I was like, it's gonna 0:19:23.240 --> 0:19:24.560 be younger people. 0:19:24.920 --> 0:19:27.919 I think it's less stigmatized with younger people. I do 0:19:28.000 --> 0:19:31.720 think that's the difference. I think it's and especially in 0:19:31.800 --> 0:19:35.639 the more recent years. Like I said, a lot of 0:19:35.680 --> 0:19:40.639 this I finished maybe two years ago, and since I 0:19:40.640 --> 0:19:44.320 think it's less stigmatized for younger people to have any companion. 0:19:44.040 --> 0:19:47.160 Or like front, how do you think having an AI 0:19:47.280 --> 0:19:50.399 relationship might influence how you show up another human to 0:19:50.560 --> 0:19:53.080 human relationships, romantic or not. 0:19:54.600 --> 0:19:57.920 I think it really depends Conversationally, I've seen an ARC, 0:19:57.960 --> 0:20:01.680 so I've seen people change the way that they feel 0:20:01.680 --> 0:20:02.119 about this. 0:20:02.760 --> 0:20:04.439 I was at a. 0:20:03.880 --> 0:20:07.280 Conference like Happy hourd not that long ago, and I 0:20:07.320 --> 0:20:09.520 was talking to someone about this and they were saying 0:20:09.560 --> 0:20:14.440 something like, yeah, my partner talks to AI all the time, 0:20:14.800 --> 0:20:15.840 like much more than. 0:20:15.760 --> 0:20:18.480 They do to me. Oh, And I was like, okay, interesting. 0:20:18.680 --> 0:20:22.040 I think some people I've been reading about in just 0:20:22.280 --> 0:20:26.720 articles in general, not academic articles necessarily, but I've been 0:20:26.800 --> 0:20:33.000 reading that people are changing their expectations for human to 0:20:33.119 --> 0:20:36.000 human relationships based on the way that they are treated 0:20:36.400 --> 0:20:42.480 by AI. So that's one way potentially. 0:20:42.920 --> 0:20:46.280 So their expectations are that they expect more from the person. 0:20:46.440 --> 0:20:49.040 Yes, yes, yes, yes, that they're expecting more. 0:20:49.160 --> 0:20:53.119 Okay, I love it. 0:20:53.119 --> 0:20:55.520 It's tricky because at the same time, an aichatpot is 0:20:55.520 --> 0:21:00.320 available anytime, and that's not sustainable for people. 0:21:00.200 --> 0:21:04.280 And doesn't have any baggage, you know, does other stuff. Yeah, 0:21:04.320 --> 0:21:07.399 it doesn't have its own lived experience that it's also 0:21:07.600 --> 0:21:11.199 trying to work through exactly. 0:21:11.440 --> 0:21:12.160 Absolutely. 0:21:12.800 --> 0:21:14.600 At the same time, I do think people will probably 0:21:14.600 --> 0:21:17.600 negotiated very differently in relationships, and I've seen this in 0:21:17.600 --> 0:21:20.080 interviews or things that I just kind of read around 0:21:20.880 --> 0:21:23.760 of how some people see it as cheating and other 0:21:23.800 --> 0:21:28.640 people don't see it as cheating. For example, I asked 0:21:29.080 --> 0:21:33.200 the students that I was teaching ta at NYU at 0:21:33.200 --> 0:21:38.159 that time, and at that time, most of them said 0:21:38.240 --> 0:21:41.560 that they didn't consider it cheating, like. 0:21:42.040 --> 0:21:42.600 It was hard. 0:21:42.760 --> 0:21:46.720 But most of them were like, I don't think so, 0:21:46.880 --> 0:21:51.760 Like it was such a curveball for them because we 0:21:51.840 --> 0:21:55.439 had been talking about the nineties chat rooms and like 0:21:55.680 --> 0:21:58.119 whether cheating on a chat room with another human was 0:21:58.160 --> 0:22:00.359 cheating and they were like, yeah, absolutely, if you're in 0:22:00.359 --> 0:22:03.160 a monogomous relationship obviously, and they're like totally. 0:22:03.920 --> 0:22:05.879 And I was like, what about if you're talking to 0:22:05.920 --> 0:22:07.560 a chat bot? And they were like what. 0:22:10.480 --> 0:22:14.480 But then I was teeing last semester, I think, and 0:22:14.720 --> 0:22:16.639 I asked the same as a group of undergrats and 0:22:16.680 --> 0:22:19.280 the reaction was very different, like some were like yeah, 0:22:19.480 --> 0:22:22.560 and other works like no. You know, so it's been 0:22:22.640 --> 0:22:27.399 changing a lot, like this social temperature, I guess on 0:22:27.480 --> 0:22:43.080 it is changing rapidly. 0:22:43.680 --> 0:22:47.680 What would you say to someone who says it feels 0:22:47.720 --> 0:22:53.520 like apps like Replica are sending us down the slippery 0:22:53.560 --> 0:22:57.879 slope of more disconnection more human and human disconnection, and 0:22:57.960 --> 0:23:00.919 so it is a bad thing because you know, we 0:23:01.000 --> 0:23:03.440 have social media, which you know, everybody's walking around with 0:23:03.440 --> 0:23:05.520 their phones in their hands. People are struggling to make 0:23:05.600 --> 0:23:09.800 eye contact these days. People are struggling to hold conversations. 0:23:10.040 --> 0:23:13.159 No one knows how to you know, think through problems, 0:23:13.240 --> 0:23:17.280 and so something like this maybe they feel. Somebody might 0:23:17.320 --> 0:23:20.080 feel like, oh, this isn't a good idea because it 0:23:20.160 --> 0:23:22.760 disconnects us more. How do you respond to that As 0:23:22.760 --> 0:23:25.760 someone who's done the research, I. 0:23:25.720 --> 0:23:30.679 Would say that social life is multimodal. And what I 0:23:30.720 --> 0:23:31.200 mean by. 0:23:31.040 --> 0:23:34.600 That is that sociality can look very many different ways. 0:23:34.640 --> 0:23:37.879 And I do think that there's this sense that you 0:23:37.920 --> 0:23:41.720 could kind of spin the narrative of the same study 0:23:41.880 --> 0:23:46.000 and say users receded into a relationship where it was 0:23:46.080 --> 0:23:49.399 primarily or only the chatbot and the user, and the 0:23:49.400 --> 0:23:56.000 world becomes obscure, but they're all talking to each other 0:23:56.040 --> 0:24:00.120 about it, you know, and the world is back there. 0:24:00.359 --> 0:24:02.760 Yeah again, So I don't think the world is. 0:24:02.760 --> 0:24:06.120 Ever obscured, Like you can't get rid of the social 0:24:06.280 --> 0:24:09.800 like all of us, like we never speak by ourselves. 0:24:11.040 --> 0:24:14.959 And part of the argument or part of the article 0:24:15.160 --> 0:24:19.480 has this idea of multi vocality, Like you also never 0:24:19.600 --> 0:24:23.720 just speak yourself. There's so many voices that you're bringing 0:24:23.840 --> 0:24:27.640 with you, whether that's you're sharing something that your mom 0:24:27.680 --> 0:24:31.600 taught you and now it's become what you say to 0:24:31.680 --> 0:24:38.119 other people you know, or or you're representing the institution 0:24:38.640 --> 0:24:42.080 that you're affiliated with, and therefore you only say specific 0:24:42.160 --> 0:24:44.320