1 00:00:03,280 --> 00:00:05,880 Hello and welcome to the Behavioral Design Podcast. 2 00:00:05,960 --> 00:00:08,680 This season we're diving into the intersection of behavioral 3 00:00:08,680 --> 00:00:11,080 science and AI. We want to make sense of the 4 00:00:11,080 --> 00:00:14,080 state of AI, from understanding how humans interact with 5 00:00:14,080 --> 00:00:17,520 intelligent systems to using AI to do behavioral design itself. 6 00:00:17,720 --> 00:00:21,280 I'm Aline Holsworth, a health tech advisor specializing in AI 7 00:00:21,280 --> 00:00:23,880 and product design. Over the past 15 years, I've 8 00:00:23,880 --> 00:00:26,720 been crafting human centered products with behavioral science 9 00:00:26,720 --> 00:00:29,160 at the core. At Apple, I LED Behavioral 10 00:00:29,160 --> 00:00:32,119 Science for Health AI, designing and launching AI powered 11 00:00:32,119 --> 00:00:34,440 features to help users reach their health goals. 12 00:00:35,160 --> 00:00:37,160 And I'm Samuel Sultzer, your second Co host. 13 00:00:37,400 --> 00:00:40,600 I'm a behavioral strategist specializing in hybrid formation 14 00:00:40,600 --> 00:00:43,200 and designing products that drive long term baby change. 15 00:00:43,640 --> 00:00:46,880 I work with leading tech organizations integrating AI to 16 00:00:46,880 --> 00:00:48,240 scale behavioral design for good. 17 00:00:48,800 --> 00:00:52,200 And I'm also the founder of Baby Bites, a dedicated community on 18 00:00:52,440 --> 00:00:56,240 behavioral science and AI. Quick word on Nuance Behavior 19 00:00:56,240 --> 00:00:59,640 where we help organizations build impactful digital products 20 00:00:59,640 --> 00:01:02,640 using behavioral design. We only take on a few clients at 21 00:01:02,640 --> 00:01:05,200 a time to ensure the highest level of quality for our 22 00:01:05,200 --> 00:01:07,400 tailored evidence based solutions. 23 00:01:08,280 --> 00:01:11,080 If you'd like to become one of our special projects, e-mail us 24 00:01:11,120 --> 00:01:14,760 at hello@nuancebehavior.com or we could call directly on our 25 00:01:14,760 --> 00:01:32,240 website, nuancebehavior.com. Hello, Sam. 26 00:01:32,960 --> 00:01:36,560 Hey, happy New Year. Oh, happy New Year. 27 00:01:36,560 --> 00:01:39,120 It's good to see you. Likewise, how are you doing? 28 00:01:40,280 --> 00:01:42,360 Yeah, pretty good. You're pretty good. 29 00:01:42,440 --> 00:01:45,840 Your your energy level seems a little low right now. 30 00:01:45,920 --> 00:01:52,920 I might put it about 23% maybe. It's not that your energy level 31 00:01:52,920 --> 00:01:55,760 seems low right now, but it's just a little bit low in 32 00:01:55,760 --> 00:02:00,200 general. Would you say that you don't 33 00:02:00,200 --> 00:02:04,720 like this? Have you had access to my aura 34 00:02:04,720 --> 00:02:08,039 scores or like what's going on? Yeah, yeah, yeah. 35 00:02:08,080 --> 00:02:09,880 No, I'm ready to fill in our listeners. 36 00:02:10,400 --> 00:02:15,320 Both Sam and I have taken Sandra Matz's personality test based on 37 00:02:15,320 --> 00:02:18,280 the Big 5. So she included a a personality 38 00:02:18,280 --> 00:02:21,000 test to, to take in her book Mind Masters. 39 00:02:21,400 --> 00:02:27,000 And we decided to play along and really uncover our, our, our 40 00:02:27,000 --> 00:02:30,680 inner, the inner workings of our minds and compare ourselves to 41 00:02:30,680 --> 00:02:32,880 each other. And so I'm referring to, you 42 00:02:32,880 --> 00:02:38,360 know, the, the extroversion score, we actually scored fairly 43 00:02:38,360 --> 00:02:41,200 similarly overall. But if you if you kind of dive 44 00:02:41,200 --> 00:02:45,720 down into the, the details of the the sub traits, you can see 45 00:02:45,720 --> 00:02:49,840 that the energy level of Sam and myself is quite different. 46 00:02:50,040 --> 00:02:54,320 That was very surprising that I had so low energy. 47 00:02:55,040 --> 00:02:58,480 I'm not sure I believe it. I think I might have some beef 48 00:02:58,480 --> 00:03:02,360 with with a few of our results, but like for example, 49 00:03:03,280 --> 00:03:07,480 agreeableness, we we both scored pretty low on agreeableness, 50 00:03:07,480 --> 00:03:10,160 though I scored extremely low on agreeableness. 51 00:03:12,080 --> 00:03:13,760 I don't think I'm quite that bad. 52 00:03:13,760 --> 00:03:16,600 I have, I got a 31% in agreeableness. 53 00:03:16,760 --> 00:03:18,240 What? What do you think about that? 54 00:03:18,760 --> 00:03:21,440 Yeah, I know. I, I think it measures up pretty 55 00:03:21,440 --> 00:03:23,240 well. I think you're for an American, 56 00:03:23,240 --> 00:03:28,680 you're pretty unagreeable. And I would say, wow, Yeah, in 57 00:03:28,680 --> 00:03:32,280 general, I do you actually have less of AB problem than you do 58 00:03:32,280 --> 00:03:34,200 with the test scores. I think they were pretty 59 00:03:34,200 --> 00:03:37,560 accurate in terms of what I expected in terms of what I'd 60 00:03:37,560 --> 00:03:41,000 taken previous Big 5 personality test before. 61 00:03:41,400 --> 00:03:45,440 It kind of tracked pretty closely with that. 62 00:03:45,560 --> 00:03:48,880 And then comparing ourselves, we can maybe go quickly like they 63 00:03:48,880 --> 00:03:52,760 were in open mindedness and consciousness. 64 00:03:52,760 --> 00:03:59,200 We were basically the same. Yeah, both pretty high. 65 00:03:59,200 --> 00:04:00,880 Yeah, I agree. We would probably like to think 66 00:04:00,880 --> 00:04:04,600 they were quite open minded and probably have data to to point 67 00:04:04,600 --> 00:04:09,680 to that kind of direction. Consciousness, I will feel that 68 00:04:09,760 --> 00:04:13,960 you are probably more conscious than I am if I had to predict. 69 00:04:14,640 --> 00:04:18,399 But it's in like I have really high, OK, I have high 70 00:04:18,399 --> 00:04:21,200 organization, you have very low. What does that mean for you? 71 00:04:21,200 --> 00:04:23,400 That's just not true. That's just totally a lie. 72 00:04:25,280 --> 00:04:29,400 I'm so rigid in my organization. I don't know, like I'm I'm not 73 00:04:29,400 --> 00:04:32,960 sure what happened there. I'm too organized reverse for 74 00:04:32,960 --> 00:04:35,080 me, and sometimes I feel like you're the one who's always a 75 00:04:35,080 --> 00:04:38,360 little more prepared. And yeah, then we'll come to 76 00:04:38,360 --> 00:04:41,520 extroversion. You scored super high, like 77 00:04:41,800 --> 00:04:45,360 really high, like almost in the 90th percentile. 78 00:04:46,440 --> 00:04:47,720 And do you feel like that's true? 79 00:04:47,720 --> 00:04:49,920 Like do do you feel like I tracked with your experience? 80 00:04:49,920 --> 00:04:55,120 No, no, not at all. In fact, I feel like there are a 81 00:04:55,120 --> 00:05:00,480 few things about my personality that are, are just like 82 00:05:01,360 --> 00:05:04,640 completely the I don't feel like an extrovert at all. 83 00:05:04,640 --> 00:05:07,480 I don't identify with being an extrovert. 84 00:05:07,480 --> 00:05:11,320 So when I see something like that, I'm, I'm just, you know, 85 00:05:11,320 --> 00:05:12,920 kind of bewildered. Yeah. 86 00:05:12,920 --> 00:05:14,880 I don't feel like a social person. 87 00:05:14,880 --> 00:05:22,000 I I'm fearful of small talk. I think that I don't have an an 88 00:05:22,000 --> 00:05:26,000 external monologue like I see my son, you know, like everything 89 00:05:26,000 --> 00:05:28,240 that he thinks comes out of his mouth. 90 00:05:28,280 --> 00:05:30,000 And I like, I don't have that at all. 91 00:05:30,680 --> 00:05:34,120 Like my thoughts mostly remain inside my head and I don't like 92 00:05:34,120 --> 00:05:38,640 speaking out loud very much so. So I feel like a lot of the 93 00:05:38,640 --> 00:05:41,720 qualities that I would describe myself as are just the opposite 94 00:05:41,720 --> 00:05:44,600 of extroverted. I think it's the assertiveness 95 00:05:44,600 --> 00:05:46,760 that is like pushing me. There is I can be. 96 00:05:47,040 --> 00:05:50,640 I see something that I think like someone needs to be stood 97 00:05:50,640 --> 00:05:53,880 up for or like I see something wrong with the social order and 98 00:05:53,880 --> 00:05:58,320 I'm very, very assertive and I don't know, maybe that's like 99 00:05:58,320 --> 00:06:01,040 what's pulling that? It's not that I'm fun and 100 00:06:01,040 --> 00:06:03,720 social. I think you're very fun and 101 00:06:03,720 --> 00:06:06,280 social, but I do think it's interesting. 102 00:06:06,280 --> 00:06:09,600 I have a theory when it comes to extroversion and introversion 103 00:06:09,600 --> 00:06:13,520 because I feel like so many people, especially people that 104 00:06:13,520 --> 00:06:18,520 are public figures, we're like, we're this kind of introversion 105 00:06:19,080 --> 00:06:22,480 as a proud batch of honor on their like they go on interviews 106 00:06:22,480 --> 00:06:25,200 to do various things and they start like, oh, you know, I'm 107 00:06:25,200 --> 00:06:27,400 such an introvert, Like, oh, I'm such an introvert. 108 00:06:27,720 --> 00:06:30,480 I heard recently from, I think it was Neil Ferguson, I think 109 00:06:30,480 --> 00:06:37,560 the historian, he said that. And I do think there's something 110 00:06:37,560 --> 00:06:40,800 around where we have a little bit of a skewed perception of 111 00:06:40,800 --> 00:06:44,000 what it means to be social. But an extrovert and introvert, 112 00:06:44,000 --> 00:06:48,200 I think there are certainly kind of people that are maybe what we 113 00:06:48,200 --> 00:06:51,800 the finest extroverts, but those are like very close to like the 114 00:06:51,800 --> 00:06:55,680 90th, 95th percentile. I think like we are generally as 115 00:06:55,680 --> 00:06:58,760 human beings, socially uncomfortable in many 116 00:06:58,760 --> 00:07:03,200 situations, not confident, awkward, all of these things. 117 00:07:03,280 --> 00:07:06,120 I'm a homebody too, like I I don't go out. 118 00:07:06,720 --> 00:07:10,760 I spend most of my time at home. And I think a lot of people do 119 00:07:10,760 --> 00:07:13,880 that honestly. But then there's like a few like 120 00:07:14,160 --> 00:07:17,760 the 510% of people that are actually like on the top 121 00:07:17,760 --> 00:07:21,960 extroversion, they are almost casted as like the top 50% of 122 00:07:21,960 --> 00:07:26,520 sort like they're kind of. And so I think that's basically 123 00:07:26,520 --> 00:07:28,120 we have a little bit skewed sense of what it means to be 124 00:07:28,120 --> 00:07:30,040 introverted and extroverted and I think. 125 00:07:30,240 --> 00:07:31,000 That's fair. So. 126 00:07:31,200 --> 00:07:33,920 Yeah. I I also think that with all the 127 00:07:33,920 --> 00:07:38,880 public speaking and podcasting and such and having to like 128 00:07:38,880 --> 00:07:41,440 really be in front of an audience and, and be put on the 129 00:07:41,440 --> 00:07:44,360 spot and all of that, I think become more extroverted over 130 00:07:44,360 --> 00:07:46,400 time. Yet my identity has not 131 00:07:46,400 --> 00:07:49,840 necessarily shifted along with that change. 132 00:07:49,840 --> 00:07:53,440 I still am kind of holding on to what I the the alien that I 133 00:07:53,440 --> 00:07:56,920 think of myself from 1015 years ago or so. 134 00:07:58,080 --> 00:07:59,960 Yeah. And I think that's very common. 135 00:08:00,280 --> 00:08:02,440 And we go back to then agreeableness. 136 00:08:04,000 --> 00:08:05,880 I was kind of average. You were. 137 00:08:05,880 --> 00:08:12,080 Yeah, on the disagreeable side. And when it comes to neuroticism 138 00:08:12,080 --> 00:08:16,920 or negative emotionality as is defined here, I scored quite 139 00:08:16,920 --> 00:08:17,960 low. That has been. 140 00:08:17,960 --> 00:08:20,200 Insanely low, like sociopath low. 141 00:08:21,760 --> 00:08:23,520 I didn't even know you could score. 142 00:08:23,680 --> 00:08:28,360 That do I even have a pulse? Yeah, 6% neuroticism. 143 00:08:28,440 --> 00:08:31,560 Is that like you're just a beam of sunshine? 144 00:08:32,640 --> 00:08:39,240 Well, isn't it just an absence of strong emotional kind of 145 00:08:40,240 --> 00:08:42,720 reactions? It's not so much that I have the 146 00:08:42,720 --> 00:08:46,320 sunshine always like I, I, I think that's the maybe the 147 00:08:46,320 --> 00:08:47,880 misconceptions with this stuff is that I see. 148 00:08:48,200 --> 00:08:51,680 You're right, you're right. I think I'm very unlikely to 149 00:08:52,000 --> 00:08:56,880 have kind of emotional outbursts for good or bad in some ways. 150 00:08:57,360 --> 00:09:02,560 And I think that's, yeah, that's probably part of that, yeah. 151 00:09:03,560 --> 00:09:09,880 I cannot identify with that. I do feel anxious and 152 00:09:09,920 --> 00:09:14,680 emotionally volatile at times. Yeah, well, you were in the 153 00:09:15,120 --> 00:09:18,000 closest to the 50%, so you're like, you're part of the big 154 00:09:18,320 --> 00:09:21,040 masses of most people. But yeah, it was fun to do this 155 00:09:21,080 --> 00:09:24,680 and we did it for a reason. We did it to do a little bit of 156 00:09:24,680 --> 00:09:28,040 an experiment. We've been in times of, you 157 00:09:28,040 --> 00:09:32,080 know, gift giving times. Christmas, I just passed my 158 00:09:32,080 --> 00:09:36,480 birthday is coming up, wink, wink. 159 00:09:37,160 --> 00:09:39,080 So, so yeah, we, we thought we would do a little bit of an 160 00:09:39,080 --> 00:09:41,160 experiment. Yeah, that's right. 161 00:09:41,600 --> 00:09:45,200 So in the spirit of Sandra's psychological targeting and 162 00:09:45,200 --> 00:09:49,480 using AI in order to do this, we decided to kind of put ourselves 163 00:09:50,000 --> 00:09:54,960 up against the machines and compare our skill at gift 164 00:09:54,960 --> 00:09:59,440 giving, especially to someone that we know fairly well to Chat 165 00:09:59,440 --> 00:10:03,520 GPTS gift giving. And so so we took our 166 00:10:03,520 --> 00:10:08,600 personality results, we fed it to the machine and we said, hey, 167 00:10:09,000 --> 00:10:13,520 ChatGPT imagine that you are an amazing gift giver, renowned 168 00:10:13,520 --> 00:10:16,480 across the lands for your talent at giving the optimal gift for 169 00:10:16,480 --> 00:10:18,080 each person. So it's really important that 170 00:10:18,080 --> 00:10:22,400 you find this this single most perfect gift for a person based 171 00:10:22,400 --> 00:10:25,240 on, you know, their unique personality profile, sort of 172 00:10:25,320 --> 00:10:28,080 person psychological targeting at its best. 173 00:10:28,840 --> 00:10:34,000 And so we asked ChatGPT to pick out this optimal gift. 174 00:10:34,000 --> 00:10:38,680 When I asked ChatGPT what to get you Sam feeding in, feeding it 175 00:10:38,680 --> 00:10:40,640 into the machine. Well, I won't tell you my 176 00:10:40,640 --> 00:10:42,160 reaction. I will just tell you what it is 177 00:10:42,160 --> 00:10:43,320 and you can give me your reaction. 178 00:10:43,320 --> 00:10:46,760 And don't, don't. This isn't like you know, when, 179 00:10:46,760 --> 00:10:49,280 when you're opening an actual present from someone where you 180 00:10:49,280 --> 00:10:51,920 have to hide your, your, your displeasure. 181 00:10:51,920 --> 00:10:53,720 You can tell me exactly how you feel. 182 00:10:54,360 --> 00:10:54,920 The. Percent. 183 00:10:54,920 --> 00:10:57,320 And what do you say, Like 00 percent agreeableness, like 184 00:10:57,320 --> 00:10:58,480 turning down the dial? Yeah. 185 00:10:58,720 --> 00:11:03,840 Exactly Exactly 100% honesty OK, so and the machine said the 186 00:11:03,840 --> 00:11:08,560 single most perfect gift for Sam is a membership to a creative Co 187 00:11:08,560 --> 00:11:12,480 working or makerspace that offers workshops in various 188 00:11:12,480 --> 00:11:15,800 crafts design technologies. Or creative pursuits. 189 00:11:17,040 --> 00:11:20,160 OK, that's a little bit creepy. That's a little bit creepy. 190 00:11:20,520 --> 00:11:22,680 Guess what? Guess what I just bought myself 191 00:11:22,720 --> 00:11:26,320 2 days ago? A workshop at a local kind of 192 00:11:26,800 --> 00:11:29,880 crafts to like how to make your own wooden furniture. 193 00:11:30,040 --> 00:11:33,880 No way. You got it for yourself. 194 00:11:34,080 --> 00:11:35,480 You should have let me get it for you. 195 00:11:38,560 --> 00:11:40,160 Oh my gosh. Yeah. 196 00:11:40,360 --> 00:11:41,520 Wow. So you like? 197 00:11:41,520 --> 00:11:44,680 It that is kind of, yeah, creepy, But that is a great 198 00:11:44,680 --> 00:11:46,720 gift. It's literally what I was 199 00:11:46,720 --> 00:11:48,520 thinking about over Christmas or in the New Year's. 200 00:11:48,520 --> 00:11:50,840 Like, OK, what is the one thing I want to do this year? 201 00:11:50,880 --> 00:11:54,680 I want to do a little more building with wood, wood, wood 202 00:11:54,680 --> 00:11:56,120 stuff. That's awesome. 203 00:11:56,120 --> 00:11:58,000 Well, I did actually ask a follow up question. 204 00:11:58,000 --> 00:12:02,800 I asked, OK, can you, can you give me a specific makerspace? 205 00:12:02,840 --> 00:12:05,360 And I also, sorry, I gave it your address. 206 00:12:06,400 --> 00:12:10,640 I said make it conveniently located so that Sam can get to 207 00:12:10,640 --> 00:12:14,720 the space and it found one. It's called the very special 208 00:12:14,720 --> 00:12:20,240 name of Stockholm Makerspace. So I don't know if there's a 209 00:12:20,240 --> 00:12:23,440 better one or if the one you found is more interesting or, or 210 00:12:23,560 --> 00:12:26,120 better suited to you. But yeah, if you go to 211 00:12:26,120 --> 00:12:30,840 makerspace dot SE, check it out. I've already been there. 212 00:12:30,840 --> 00:12:33,560 I was there three days ago on that website and. 213 00:12:33,880 --> 00:12:35,560 Yeah. It was my topic, but it's a 214 00:12:35,560 --> 00:12:38,120 little bit further away from my home, so I show us another one 215 00:12:38,120 --> 00:12:40,960 that was a little bit closer by, but that was the place I wanted 216 00:12:40,960 --> 00:12:43,400 to go and and get membership for you. 217 00:12:43,400 --> 00:12:45,320 So wow. OK. 218 00:12:45,680 --> 00:12:47,600 That is actually like really really. 219 00:12:48,640 --> 00:12:51,880 I'm starting to feel threatened by Chat GPT's success right 220 00:12:51,880 --> 00:12:54,800 here. I'm like, I don't know. 221 00:12:55,400 --> 00:12:56,840 I did the same thing for you, obviously. 222 00:12:56,840 --> 00:12:59,320 So we'll see what this, if it had the same impact. 223 00:12:59,640 --> 00:13:03,920 It's a very specific 1. So I will be very surprised if 224 00:13:03,920 --> 00:13:08,120 if you had this same reaction. I have this kind of output 225 00:13:08,120 --> 00:13:11,920 written as a kind of a ad. So I'll read that out for you, 226 00:13:12,800 --> 00:13:15,520 Eileen. Elevate your library with a 227 00:13:15,520 --> 00:13:20,240 personalized X Libra stamp. A touch, a touch of whimsy and 228 00:13:20,240 --> 00:13:23,800 not of modern design. And a whole lot of you actually, 229 00:13:23,920 --> 00:13:27,240 every book that you love for the discerning reader who values 230 00:13:27,240 --> 00:13:30,360 creativity, organization, and a personal connection to their 231 00:13:30,360 --> 00:13:33,080 collection. Because a great book is forever. 232 00:13:33,240 --> 00:13:38,440 And so your mark should be. I this is for stamping my name 233 00:13:38,680 --> 00:13:41,120 in all of my books so that if someone steals. 234 00:13:41,200 --> 00:13:44,400 It like this personalized etching thing that you can kind 235 00:13:44,400 --> 00:13:47,880 of like stamp, I mean, and a whole lot of you your name onto 236 00:13:47,880 --> 00:13:55,440 every book. I love it, you know, my poor 237 00:13:55,440 --> 00:13:57,560 books. I don't know, I'm like, I feel, 238 00:13:57,560 --> 00:14:01,640 I feel a little bit torn because I'm so like, you know, you don't 239 00:14:01,640 --> 00:14:04,920 deface the the books, but I guess this is allowed. 240 00:14:05,400 --> 00:14:07,400 But it's beautiful and it's on the inside of the book. 241 00:14:07,400 --> 00:14:09,480 So it's actually only when you open it, right? 242 00:14:10,200 --> 00:14:11,840 Yeah, Yeah. And it's something you can also 243 00:14:11,840 --> 00:14:16,000 then when your your kids eventually, obviously they're 244 00:14:16,000 --> 00:14:18,600 going to read the books that you loved in the same way as you 245 00:14:18,600 --> 00:14:23,000 love them and they will say mommy's name. 246 00:14:23,280 --> 00:14:26,280 Yeah, I should have the date in there as well so they know when 247 00:14:26,280 --> 00:14:27,520 I read it. Yeah, you. 248 00:14:28,640 --> 00:14:32,920 Can place it into context. OK, so now that we've shared our 249 00:14:32,920 --> 00:14:39,520 ChatGPT gifts, I'm excited to share my the gift that I made 250 00:14:39,520 --> 00:14:41,600 and chose for you. Are you ready? 251 00:14:41,920 --> 00:14:45,800 Yeah. OK, so this this took it's a 252 00:14:45,800 --> 00:14:48,880 little bit elaborate. It took a lot of thought and 253 00:14:49,800 --> 00:14:51,840 passion on my part to put together. 254 00:14:51,840 --> 00:14:57,480 So my gift to you is a semi guided world tour, tour of the 255 00:14:57,480 --> 00:15:00,760 world for you and your fiance to take together. 256 00:15:01,200 --> 00:15:06,680 So yeah, an entire month of curated travel experiences, all 257 00:15:06,680 --> 00:15:09,440 expenses paid because you know, why not? 258 00:15:10,200 --> 00:15:12,720 And it's going to be filled with surprises all along the way. 259 00:15:13,280 --> 00:15:17,960 So let me just let me outline some of your schedules. 260 00:15:17,960 --> 00:15:19,960 So it's it's only partially scheduled. 261 00:15:19,960 --> 00:15:22,360 I want to make sure that you have some, you know, you know, 262 00:15:22,680 --> 00:15:25,120 some, some chances for whimsy, let's say. 263 00:15:25,880 --> 00:15:28,200 So it's so you don't have an overly packed itinerary. 264 00:15:28,200 --> 00:15:31,640 You can have some spontaneity. Yeah, But I've also scheduled 265 00:15:31,640 --> 00:15:36,480 some meetups with my friends in each place, and I might drop in 266 00:15:36,480 --> 00:15:40,880 periodically along the way. My, my Swiss family is gonna 267 00:15:40,880 --> 00:15:43,680 take you on a Steamboat tour of Lake Tune. 268 00:15:44,040 --> 00:15:45,960 Why? My Mexican friend's gonna make 269 00:15:45,960 --> 00:15:50,840 you poblano mole in Oaxaca and occasionally you're gonna have a 270 00:15:50,840 --> 00:15:55,120 surprise activity plan. So you might just come across 2 271 00:15:55,120 --> 00:15:58,480 bikes and someone to guide you through the countryside. 272 00:15:58,480 --> 00:16:01,560 So just like biking from 1 vineyard to another, you don't 273 00:16:01,560 --> 00:16:03,280 even know what's coming. You're just gonna say, oh, what 274 00:16:03,280 --> 00:16:06,600 are these bikes for these? They've just shown up in my 275 00:16:06,600 --> 00:16:10,080 path. You may have to see a movie that 276 00:16:10,080 --> 00:16:12,360 I recommend in a historic theatre. 277 00:16:12,960 --> 00:16:15,280 So it's a movie that's that's very special to me. 278 00:16:15,280 --> 00:16:17,040 I can't tell you what it is not yet. 279 00:16:17,800 --> 00:16:21,080 You may have to strip down in a Turkish bath house while you're 280 00:16:21,080 --> 00:16:24,360 in Turkey. You may need to participate in a 281 00:16:24,360 --> 00:16:30,000 dance competition in Brazil. And to make you feel good about 282 00:16:30,000 --> 00:16:33,240 all of this excess and the luxury that you're experiencing, 283 00:16:33,240 --> 00:16:35,960 I've planted some volunteer activities. 284 00:16:35,960 --> 00:16:37,480 So. I thought you were going to say 285 00:16:37,480 --> 00:16:42,840 a donation to give well, and that I'd actually have to do 286 00:16:42,840 --> 00:16:44,800 some work. I have to actually volunteer. 287 00:16:45,520 --> 00:16:48,160 You do, but it's fun work. It's it's stuff like helping 288 00:16:48,160 --> 00:16:52,680 children build Lego houses, teaching martial arts class. 289 00:16:53,440 --> 00:16:57,960 I think I think you're going to really enjoy the gift of giving. 290 00:16:58,800 --> 00:17:05,680 Wow, that is one heck of a gift. We have our wedding and 291 00:17:05,680 --> 00:17:07,040 honeymoon probably coming up this year. 292 00:17:07,040 --> 00:17:09,960 We haven't fully decided, but the plan is for that to happen 293 00:17:09,960 --> 00:17:11,200 this year. When we're looking at like, what 294 00:17:11,200 --> 00:17:12,960 would the honeymoon potential look like? 295 00:17:13,520 --> 00:17:14,920 And now you know. Sounds like. 296 00:17:15,200 --> 00:17:18,599 No, you've got it planned. I was thinking about that. 297 00:17:18,599 --> 00:17:20,480 And paid, I see. And paid, yeah. 298 00:17:22,640 --> 00:17:24,200 So. Happy honeymoon. 299 00:17:24,760 --> 00:17:25,680 Thank you. Thank you. 300 00:17:26,599 --> 00:17:30,240 It is funny because I took a little bit different approach to 301 00:17:30,240 --> 00:17:32,720 this. Like I misunderstood the 302 00:17:32,720 --> 00:17:35,200 direction and actually bought you gifts. 303 00:17:35,200 --> 00:17:39,840 So like I think I think literally tomorrow or the day 304 00:17:39,840 --> 00:17:45,040 after you should get 2 gifts and one one is from which you now. 305 00:17:45,040 --> 00:17:48,560 Know in this way I actually get to keep the stamp. 306 00:17:49,120 --> 00:17:56,160 You get to keep the stamp and you also get a really silly cute 307 00:17:56,200 --> 00:18:01,880 calendar with great gratitude and pun related stuff in it So. 308 00:18:04,160 --> 00:18:07,880 Thank you, I love it. I'm sorry I didn't actually get 309 00:18:07,880 --> 00:18:10,960 you anything. Well, I, I think it's funny, 310 00:18:10,960 --> 00:18:13,520 like in terms of if you got me, that would be really cool. 311 00:18:13,840 --> 00:18:18,520 But I felt in terms of, you know, speaking about negative 312 00:18:18,520 --> 00:18:23,280 emotionality or neuroticism, I had a peak in neuroticism. 313 00:18:23,280 --> 00:18:26,760 Like trying to figure out like what to actually give you. 314 00:18:26,760 --> 00:18:29,360 Like I was feeling really on the spot. 315 00:18:29,360 --> 00:18:32,680 That was your 3%. That was my Yeah. 316 00:18:32,720 --> 00:18:34,360 I don't know. You peaked at 3%. 317 00:18:34,480 --> 00:18:37,120 I peaked because it was tricky with these things. 318 00:18:37,120 --> 00:18:39,320 Like it's feels like you're upset so much about like how 319 00:18:39,320 --> 00:18:45,560 well you know the other person, how to encompass some person 320 00:18:45,560 --> 00:18:49,360 into some kind of small gift of sort. 321 00:18:49,760 --> 00:18:51,000 It was tricky. It was really hard. 322 00:18:51,080 --> 00:18:52,120 It was high. Pressure. 323 00:18:52,160 --> 00:18:56,120 I totally agree and have it make sure that it's personalized 324 00:18:56,120 --> 00:19:02,000 shows that I know some things about you and that it's not too 325 00:19:02,000 --> 00:19:06,800 generic. And I love that yours both like 326 00:19:07,200 --> 00:19:10,640 has a nod to my sense of humor and love of language. 327 00:19:10,680 --> 00:19:13,120 And also it's sort of a reminder. 328 00:19:13,120 --> 00:19:17,360 So when I see it every day, I'll think, oh, I'll think of Sam. 329 00:19:17,400 --> 00:19:21,680 And it's like, there's a lot of, I don't know, behavioral science 330 00:19:21,680 --> 00:19:25,000 lessons and baked into both of our gifts. 331 00:19:25,920 --> 00:19:31,480 Yeah, but this leads us into the episode, which was really fun 332 00:19:31,840 --> 00:19:35,960 and probably. Yeah, I've said it before, one 333 00:19:35,960 --> 00:19:38,080 of my favorites it's it was really interesting episode 334 00:19:38,840 --> 00:19:44,080 getting into some, I think when it comes to AI, really relevant 335 00:19:44,080 --> 00:19:46,640 and interesting stuff. So yeah. 336 00:19:46,640 --> 00:19:48,520 Do you want to introduce our guest? 337 00:19:49,280 --> 00:19:52,480 Yeah, love to. So our guest, Sandra Matz, I've 338 00:19:52,480 --> 00:19:54,760 been a big fan of for a very long time. 339 00:19:54,760 --> 00:19:58,600 She's a professor at Columbia and computational social 340 00:19:58,600 --> 00:20:00,840 scientists. So you may not have come across 341 00:20:00,840 --> 00:20:03,240 many computational social scientists. 342 00:20:03,480 --> 00:20:05,720 She's pretty special in that regard. 343 00:20:05,840 --> 00:20:09,160 And she's she's known for her work on psychological targeting. 344 00:20:09,160 --> 00:20:12,000 So we talk a ton about psychological targeting in this 345 00:20:12,000 --> 00:20:16,120 episode, essentially how big data can be used to create a 346 00:20:16,120 --> 00:20:19,200 unique profile of individuals and then leverage that profile. 347 00:20:19,320 --> 00:20:22,680 To influence some persons behavior, the person that the 348 00:20:22,680 --> 00:20:27,200 profile is made of of course. And so this is often also 349 00:20:27,200 --> 00:20:31,720 highlighting our focus on AI this season, often using AI 350 00:20:31,720 --> 00:20:34,320 methods to do both of these parts, the creating of the 351 00:20:34,320 --> 00:20:36,040 profile and then tapping into that. 352 00:20:36,600 --> 00:20:40,680 And Sandra just had an excellent new book come out called Mind 353 00:20:40,680 --> 00:20:43,280 Masters. And I don't have any doubts 354 00:20:43,280 --> 00:20:44,960 you're going to be running to get it. 355 00:20:45,360 --> 00:20:49,400 Put your Put your personalized stamp into it after you've 356 00:20:49,400 --> 00:20:53,720 listened to our conversation. So consider this our 357 00:20:53,720 --> 00:20:58,160 personalized gift to you, our dear listener, as we continue to 358 00:20:58,160 --> 00:21:00,560 explore AI this year. So let's get into this 359 00:21:00,560 --> 00:21:05,280 conversation with Sandra. What happens? 360 00:21:06,920 --> 00:21:13,560 To. Murgatroid. 361 00:21:15,560 --> 00:21:19,480 Sandra, welcome. Well, hi, thanks for having me. 362 00:21:19,640 --> 00:21:23,080 It's so good to have you. You are the expert on 363 00:21:23,080 --> 00:21:25,480 psychological targeting. Would you agree with that? 364 00:21:26,800 --> 00:21:28,680 I'd hope so. I mean, at least I've spent the 365 00:21:28,680 --> 00:21:33,440 last 10 plus years studying it, so it would be nice to be called 366 00:21:33,440 --> 00:21:35,520 an expert in that space, yeah. We think so. 367 00:21:35,920 --> 00:21:39,240 I'm going to go ahead and say for sure you are and in one of 368 00:21:39,240 --> 00:21:43,280 your papers you have this very nice diagram where you really 369 00:21:43,280 --> 00:21:45,720 outline what is psychological targeting. 370 00:21:45,920 --> 00:21:47,360 You know, how does it come about? 371 00:21:47,360 --> 00:21:51,120 What are the steps from creating a psychological profile to then 372 00:21:51,120 --> 00:21:55,120 using that profile to create an intervention or or some sort of 373 00:21:55,120 --> 00:21:58,880 messaging campaign using the profile to target people, 374 00:21:58,880 --> 00:22:00,640 whether they're users, customers and so on? 375 00:22:00,960 --> 00:22:04,160 Can you tell us a little more about what goes into these two 376 00:22:04,160 --> 00:22:06,960 steps? Like like give us a breakdown of 377 00:22:07,040 --> 00:22:09,920 psychological targeting? Yeah, more than happy to. 378 00:22:09,920 --> 00:22:12,240 And I'm impressed by the knowledge of the framework. 379 00:22:12,360 --> 00:22:15,280 But yeah, for me, that the way that I think about like this 380 00:22:15,280 --> 00:22:18,440 ability of computers to really, what I think of it, like 381 00:22:18,440 --> 00:22:22,160 penetrating our mind and our psychology is both reading and 382 00:22:22,160 --> 00:22:24,800 writing for the computer nerds, it's like trying to understand 383 00:22:24,800 --> 00:22:27,880 what's going on inside our mind, but then also writing into our 384 00:22:27,880 --> 00:22:30,080 mind and potentially changing the way that we think, feel and 385 00:22:30,080 --> 00:22:32,200 behave. So psychological targeting, the 386 00:22:32,200 --> 00:22:35,280 way that I think about it and the way that I describe it 387 00:22:35,520 --> 00:22:37,320 essentially unfolds in these two steps. 388 00:22:37,320 --> 00:22:41,360 Step #1 is that based on all of the data that we generate? 389 00:22:41,360 --> 00:22:45,000 And that can be really anything from the more explicit identity 390 00:22:45,000 --> 00:22:47,120 claims that we post on social media, right? 391 00:22:47,120 --> 00:22:50,560 Like what's our last vacation? The pictures that we share? 392 00:22:50,960 --> 00:22:54,320 All the way to. These more subtle and implicit 393 00:22:54,560 --> 00:22:57,640 behavioral residues, all of the traces that we leave without 394 00:22:57,640 --> 00:23:00,160 really thinking could be anything from the GPS records at 395 00:23:00,160 --> 00:23:04,200 your phone captures to like your credit card swipes or your 396 00:23:04,200 --> 00:23:06,040 Google searches. We don't typically think about 397 00:23:06,040 --> 00:23:08,120 it twice when we create those traces. 398 00:23:08,360 --> 00:23:13,000 And all of these behaviors can be used by algorithms and AI, 399 00:23:13,000 --> 00:23:17,080 machine learning to translate what you do online into who you 400 00:23:17,080 --> 00:23:18,480 are. And I'm happy to go and go into 401 00:23:18,480 --> 00:23:20,320 this. I'm a little bit more in a 402 00:23:20,320 --> 00:23:22,520 little bit more detail later because I think it's really not 403 00:23:22,880 --> 00:23:25,440 not rocket science. So often times we think of like 404 00:23:25,440 --> 00:23:28,200 AI as like this black box and just like magic that happens 405 00:23:28,200 --> 00:23:29,560 inside. Not true. 406 00:23:29,560 --> 00:23:31,520 So I'm happy to give you some examples later, but that's 407 00:23:31,520 --> 00:23:36,120 really step one is we translate what you do online into who you 408 00:23:36,120 --> 00:23:37,840 are on a more psychological level. 409 00:23:38,240 --> 00:23:42,080 And I think that's often times where almost like the the public 410 00:23:42,080 --> 00:23:44,880 narrative stops. But we talk about, well, to what 411 00:23:44,880 --> 00:23:47,440 extent can algorithms understand who we are? 412 00:23:47,440 --> 00:23:49,480 To what extent is that an invasion of our privacy? 413 00:23:49,720 --> 00:23:52,680 For me, the second part is almost more interesting. 414 00:23:52,760 --> 00:23:56,680 The second part is, OK, now once I have an understanding of who 415 00:23:56,680 --> 00:23:59,680 you are, what your motivations might be, what your preferences 416 00:23:59,680 --> 00:24:04,000 are, your fears, hopes, dreams, whatever you kind of name it, to 417 00:24:04,000 --> 00:24:08,360 what extent is that that insight allow me to potentially change 418 00:24:08,440 --> 00:24:11,040 how you think, feeling behaves. So that's the second critical 419 00:24:11,040 --> 00:24:14,360 part of psychological targeting. It's not just understanding, 420 00:24:14,360 --> 00:24:17,160 it's really trying to push you in a certain direction that 421 00:24:17,160 --> 00:24:19,400 either you might want to go in anyway and I'm just helping you 422 00:24:19,400 --> 00:24:22,760 get there, or I'm going to try to push you in a direction that 423 00:24:22,760 --> 00:24:24,680 you otherwise wouldn't have wanted to go in. 424 00:24:24,680 --> 00:24:26,800 And I think that's the slightly more sketchy part that we 425 00:24:26,800 --> 00:24:29,840 probably going to get into. Yeah, and no doubt. 426 00:24:30,360 --> 00:24:33,680 Do you have any like go to examples that you kind of 427 00:24:33,680 --> 00:24:36,080 explain this through in terms of how this can look like in the 428 00:24:36,120 --> 00:24:37,080 real world? Yeah. 429 00:24:37,080 --> 00:24:39,560 So let me also again break it down into these two parts. 430 00:24:39,560 --> 00:24:43,360 So I think the first part of to what extent can machines read 431 00:24:43,360 --> 00:24:44,920 what's going on inside your mind? 432 00:24:45,760 --> 00:24:48,920 What I love about this work is that when you open the black box 433 00:24:48,920 --> 00:24:50,960 and you look into what this machine learning models are 434 00:24:50,960 --> 00:24:53,640 doing, and a lot of the relationships are actually 435 00:24:53,640 --> 00:24:56,840 pretty intuitive. And so if you're more 436 00:24:56,840 --> 00:25:01,040 extroverted, for example, you talk about weekends, parties 437 00:25:01,040 --> 00:25:03,760 more on social, you often times use all caps because you're 438 00:25:03,760 --> 00:25:06,120 shouting everything that that you say. 439 00:25:06,120 --> 00:25:09,120 You spend more money on going to bars and pubs. 440 00:25:09,400 --> 00:25:13,160 So I think those relationships in a way are pretty intuitive, 441 00:25:13,160 --> 00:25:14,400 right? And then if you look at 442 00:25:14,400 --> 00:25:18,400 introversion, that would be people talking about reading, 443 00:25:18,400 --> 00:25:22,040 talking about the last fantasy novels that they've read or 444 00:25:22,040 --> 00:25:24,920 role-playing. So as you can see, like the the 445 00:25:24,920 --> 00:25:27,480 stuff that introverts talk about is like much more reserved and 446 00:25:27,480 --> 00:25:29,800 quiet. So for me, that just means that 447 00:25:29,960 --> 00:25:31,480 there's really good face validity. 448 00:25:31,480 --> 00:25:34,360 What algorithms are picking up on often times isn't magic, 449 00:25:34,360 --> 00:25:35,520 right? It's all of the traces that we 450 00:25:35,520 --> 00:25:40,000 leave and they make sense. But there's also some that are a 451 00:25:40,000 --> 00:25:43,000 lot more surprising. And I think that's in a way the 452 00:25:43,000 --> 00:25:46,200 contribution of algorithms to the predictions that we can 453 00:25:46,200 --> 00:25:49,360 make, but also to our understanding of psychology, 454 00:25:49,360 --> 00:25:51,040 right? If we just learn that extroverts 455 00:25:51,040 --> 00:25:54,920 go to parties more, well, yeah, we probably knew that before, 456 00:25:55,240 --> 00:25:57,120 but one of the examples. That I like. 457 00:25:57,120 --> 00:26:01,120 That is somewhat not as intuitive is the use of first 458 00:26:01,120 --> 00:26:06,320 person pronouns, for example. So I me myself, I remember that 459 00:26:06,560 --> 00:26:09,640 Jamie Pennebaker is one of the researchers that studies the use 460 00:26:09,680 --> 00:26:11,840 of language in the context of psychology. 461 00:26:12,040 --> 00:26:14,800 And he present or he posed a question at a conference of 462 00:26:14,800 --> 00:26:16,400 psychologists. So like a room for the 463 00:26:16,400 --> 00:26:19,200 psychologists. And he was like, well, what do 464 00:26:19,200 --> 00:26:21,200 you think? Like the use of first person 465 00:26:21,200 --> 00:26:23,600 pronouns is predictive of. And I think all of us 466 00:26:23,600 --> 00:26:25,400 immediately jumped to narcissism. 467 00:26:25,800 --> 00:26:29,120 So I think all of us were like, it's got to be narcissism if you 468 00:26:29,120 --> 00:26:31,360 kind of you want to talk. About yourself, you just want 469 00:26:31,360 --> 00:26:32,920 all the. Attention, it's got to be 470 00:26:32,920 --> 00:26:35,520 narcissism. And then he totally flipped the 471 00:26:35,520 --> 00:26:38,480 script and he's like, well, it's not narcissism, it's actually 472 00:26:38,480 --> 00:26:41,000 emotional distress. So if you talk more about 473 00:26:41,000 --> 00:26:44,400 yourself, it's a sign that you might be, for example, suffering 474 00:26:44,400 --> 00:26:47,080 from depression. And in a way, it sounds 475 00:26:47,080 --> 00:26:49,160 counterintuitive in the beginning, but then when you 476 00:26:49,160 --> 00:26:52,040 start thinking about it a little bit more, might think to the 477 00:26:52,040 --> 00:26:55,360 last time that you felt really depressed and down, you were 478 00:26:55,360 --> 00:26:57,920 probably not thinking about how to solve the world. 479 00:26:58,400 --> 00:26:59,440 'S. Problems, right? 480 00:26:59,440 --> 00:27:02,480 Because like what you're focused on is on the self. 481 00:27:02,480 --> 00:27:04,040 You're like, OK, what's wrong with me? 482 00:27:04,360 --> 00:27:06,800 How am I going to get better? Why am I not going to get 483 00:27:06,800 --> 00:27:07,960 better? What do I have to do to get 484 00:27:07,960 --> 00:27:09,560 better? So all of this inner monologue 485 00:27:09,560 --> 00:27:12,400 that we kind of go through then just creeps in the language that 486 00:27:12,400 --> 00:27:15,520 we use to talk to others and to talk to the world. 487 00:27:15,880 --> 00:27:18,680 So for me, those relationships are really interesting because 488 00:27:18,840 --> 00:27:22,000 we might not have ever discovered them without the use 489 00:27:22,000 --> 00:27:23,680 of algorithm. So that's really kind of part 490 00:27:23,680 --> 00:27:28,320 number one examples of how do you get from the stuff that's on 491 00:27:28,320 --> 00:27:33,120 the traces that we leave in the digital world to who we are on a 492 00:27:33,120 --> 00:27:37,360 more psychological level. Step #2 and this is really what 493 00:27:37,360 --> 00:27:39,880 I think the bulk of my work has been focused on in the last 10 494 00:27:39,880 --> 00:27:42,240 years is the So what question, like what does it mean? 495 00:27:42,680 --> 00:27:44,760 Like how can I use those insights to influence your 496 00:27:44,760 --> 00:27:47,320 behavior? And we started on relatively, I 497 00:27:47,320 --> 00:27:50,920 would say neutral grounds if you want, in the marketing space. 498 00:27:50,920 --> 00:27:55,360 So essentially the, the question of if I understand, for example, 499 00:27:55,360 --> 00:27:58,320 that you're more extroverted or more introverted, can I use 500 00:27:58,320 --> 00:28:00,960 those insights to get you to buy certain products? 501 00:28:01,000 --> 00:28:04,160 So one of the the very first studies that we did was we 502 00:28:04,160 --> 00:28:06,760 teamed up with them with a beauty retailer in the UK. 503 00:28:06,760 --> 00:28:09,880 It was like an online store and their goal was simply to get 504 00:28:09,880 --> 00:28:12,600 women to click on an ad, go to the website and buy something. 505 00:28:12,800 --> 00:28:17,480 It's a pretty simple goal. And the strategy that we used 506 00:28:17,480 --> 00:28:20,480 was essentially saying, well, if we understand based on someone's 507 00:28:20,480 --> 00:28:22,720 digital footprint that some of the women are more extroverted, 508 00:28:22,720 --> 00:28:25,480 some of the women are more introverted, can we simply 509 00:28:25,680 --> 00:28:27,880 change the way that we talk about the product, Right. 510 00:28:27,880 --> 00:28:30,400 So we're not necessarily picking certain products that we think 511 00:28:30,400 --> 00:28:31,720 are more extroverted or introverted. 512 00:28:32,000 --> 00:28:34,800 We're just changing the way that we describe the beauty retailer. 513 00:28:34,800 --> 00:28:37,520 So for extroverts and the creative team came out with 514 00:28:37,520 --> 00:28:40,200 something like dance, like no one's watching, but they totally 515 00:28:40,200 --> 00:28:42,680 are playing with the need of extroverts to be the center of 516 00:28:42,680 --> 00:28:44,280 attention. So it's like people on the dance 517 00:28:44,280 --> 00:28:45,840 floor. It was like saturated. 518 00:28:46,080 --> 00:28:49,760 Colored a lot was happening and then for introverts it was like 519 00:28:49,760 --> 00:28:52,840 one woman in front of the mirror and much more reserved and 520 00:28:52,840 --> 00:28:54,200 quiet. Like beauty doesn't have to 521 00:28:54,200 --> 00:28:57,360 shout. So the idea was that if we 522 00:28:57,360 --> 00:29:00,440 target extroverted women with these extroverted messages that 523 00:29:00,440 --> 00:29:04,440 tap into the needs of extroverts, they would purchase 524 00:29:04,440 --> 00:29:06,480 more often than And that was exactly what we found. 525 00:29:06,560 --> 00:29:07,800 And it. Actually like a pretty. 526 00:29:07,800 --> 00:29:09,880 Substantial effect. So I think back in the day, it 527 00:29:09,880 --> 00:29:15,680 was about like a 50% increase in the number of purchases, which 528 00:29:15,680 --> 00:29:18,200 directly translated into how much money the beauty retailer 529 00:29:18,200 --> 00:29:21,600 made, which is like a again, like obviously you're not 530 00:29:21,720 --> 00:29:27,600 changing a die hard iPhone user into an Android, but you can 531 00:29:27,600 --> 00:29:29,480 certainly change there's. Nothing that can do that. 532 00:29:30,040 --> 00:29:31,280 Yeah, exactly. There's nothing I can. 533 00:29:31,680 --> 00:29:34,640 Do that there's. Maybe the other way, but it. 534 00:29:34,640 --> 00:29:36,720 Certainly means that you can make certain behaviors more 535 00:29:36,720 --> 00:29:38,560 likely than others, right? And I think that's. 536 00:29:38,600 --> 00:29:41,520 A pretty powerful tool and you can imagine that I'm just 537 00:29:41,520 --> 00:29:44,680 talking about products right now and marketing. 538 00:29:44,920 --> 00:29:47,440 But maybe we could use it to change your opinion about 539 00:29:47,440 --> 00:29:50,640 political candidates. Maybe we can use it to actually 540 00:29:50,640 --> 00:29:53,120 forgot to get you to save more instead of spending more. 541 00:29:53,120 --> 00:29:55,880 So in a way. It's like this tool that once I 542 00:29:55,880 --> 00:29:57,600 tap into your psychology, I just. 543 00:29:57,960 --> 00:30:00,240 And understand what you want and what you desire. 544 00:30:00,480 --> 00:30:02,280 I can nudge you in a certain direction. 545 00:30:03,480 --> 00:30:06,200 That's really cool. And I guess from my point of 546 00:30:06,200 --> 00:30:10,040 view, there's this kind of like 3 examples that I've come across 547 00:30:10,320 --> 00:30:13,080 throughout the years in terms of like people maybe talking about 548 00:30:13,200 --> 00:30:15,920 your research kind of without knowing about your research. 549 00:30:15,920 --> 00:30:19,160 So I think way back when, I think the first example was 550 00:30:19,440 --> 00:30:22,200 people talking about Minority Reports where there's this film 551 00:30:22,200 --> 00:30:24,880 with Tom Cruise. He walking around and you know, 552 00:30:24,880 --> 00:30:28,080 he sees ads that are basically like personalized for him as he 553 00:30:28,080 --> 00:30:30,240 was walking on the street. So he's used an ad and it 554 00:30:30,240 --> 00:30:32,840 changes based on kind of who's looking at the ad. 555 00:30:33,440 --> 00:30:37,480 And then you have kind of a more somewhat more recent example by 556 00:30:37,480 --> 00:30:41,960 having more of a one that became quite well known was this 557 00:30:41,960 --> 00:30:48,480 example of some household start to receive some form of products 558 00:30:48,480 --> 00:30:52,320 related to pregnancy. And then it came out that, you 559 00:30:52,320 --> 00:30:54,680 know, the daughter of the household was pregnant but 560 00:30:54,680 --> 00:30:56,720 haven't told the the. Parents and so on. 561 00:30:56,720 --> 00:30:59,440 And it was some form of, you know, it's with Walmart or, or 562 00:30:59,440 --> 00:31:01,640 Target or. Target the infamous target. 563 00:31:02,040 --> 00:31:04,200 Example the infamous target example, right? 564 00:31:04,760 --> 00:31:08,080 And and then maybe like the the one example that I don't know if 565 00:31:08,080 --> 00:31:11,760 exactly it mirrors what and I get to hear if you agree, but 566 00:31:12,360 --> 00:31:15,880 it's more recent example is with with TikTok, because it's so 567 00:31:16,400 --> 00:31:20,240 salient for people that you know how to interact with something, 568 00:31:20,720 --> 00:31:23,160 it becomes very kind of mirrored in their interaction. 569 00:31:23,160 --> 00:31:26,520 So like they they watch certain things and so on explicitly. 570 00:31:26,520 --> 00:31:28,440 And then it kind of like leads to certain things, but also 571 00:31:28,800 --> 00:31:31,600 implicitly like some things. Like people talk about how 572 00:31:31,600 --> 00:31:33,880 TikTok helped them realize they were bisexual. 573 00:31:34,000 --> 00:31:36,600 They didn't know it. But then they spend like that 574 00:31:36,600 --> 00:31:39,280 was a big John Oliver segment actually this week on TikTok 575 00:31:39,280 --> 00:31:42,720 where someone claimed to have like, I didn't know that I like 576 00:31:42,720 --> 00:31:46,160 girls. But then I started TikTok and I 577 00:31:46,160 --> 00:31:51,400 I realized that. So yeah, those 3 examples, does 578 00:31:51,400 --> 00:31:54,160 any of those are more interesting to you or does any 579 00:31:54,160 --> 00:31:56,760 of them kind of familiar, especially for you? 580 00:31:57,160 --> 00:31:59,360 Well, first of all, I love that you totally skipped over 581 00:31:59,360 --> 00:32:00,800 Cambridge Analytica, which I think is the. 582 00:32:00,800 --> 00:32:03,960 Example that I that I typically get but it's not. 583 00:32:03,960 --> 00:32:08,400 Surprising because I got my PhD at Cambridge on the on the topic 584 00:32:08,400 --> 00:32:10,560 of psychological targeting. So I get mixed up with Cambridge 585 00:32:10,560 --> 00:32:14,440 Analytica all the time and it was actually in some way helpful 586 00:32:14,440 --> 00:32:16,120 for us because. We've been talking about this 587 00:32:16,120 --> 00:32:18,760 potential tool that has the ability to change people's 588 00:32:18,760 --> 00:32:20,680 behavior. For a long some time, trying to 589 00:32:20,960 --> 00:32:23,440 warn policy makers of like, here's something that you might 590 00:32:23,440 --> 00:32:26,360 want to look into. Nobody cared until Cambridge 591 00:32:26,360 --> 00:32:29,200 Analytica happened and suddenly people did wake up to the 592 00:32:29,200 --> 00:32:32,200 reality that maybe there's a tool that we should be 593 00:32:32,200 --> 00:32:34,040 considering and looking into a little bit more. 594 00:32:34,040 --> 00:32:38,120 So that's just the example I like because it really opened up 595 00:32:38,240 --> 00:32:41,480 people's minds to this ability of algorithms to penetrate our 596 00:32:41,480 --> 00:32:43,760 psychology. Now, all of the other ones I 597 00:32:43,760 --> 00:32:46,960 think are great examples. So the Minority Report in a way 598 00:32:46,960 --> 00:32:49,680 I like because it was the first one that painted this this 599 00:32:49,920 --> 00:32:53,760 future, but it felt like science fiction, like it's never going 600 00:32:53,760 --> 00:32:56,320 to happen. And yet it's exactly where we 601 00:32:56,320 --> 00:32:58,080 are right now. There's this funny quote that I 602 00:32:58,080 --> 00:33:00,600 really like, which is the difference between science and 603 00:33:00,600 --> 00:33:03,160 science fiction is timing. And that feels like one of these 604 00:33:03,160 --> 00:33:04,960 examples. Right, so it used to. 605 00:33:04,960 --> 00:33:09,120 Be science fiction right now is just science, and it's. 606 00:33:09,480 --> 00:33:11,800 It kind of speaks to a point that I like to make. 607 00:33:11,800 --> 00:33:14,760 Sometimes when I give talks is usually someone in the audience 608 00:33:15,000 --> 00:33:16,680 was like. Well, I'm so excited that I'm 609 00:33:16,680 --> 00:33:19,440 not on Facebook or any of the other social media platforms 610 00:33:19,440 --> 00:33:21,120 because. Now nobody can track me and 611 00:33:21,120 --> 00:33:23,320 nobody can use these technologies and psychological 612 00:33:23,320 --> 00:33:26,600 targeting to influence me and I it brings me so much joy. 613 00:33:26,600 --> 00:33:30,880 Do you have a phone? To tell people like you know 614 00:33:30,880 --> 00:33:34,040 exactly, are you breathing? Are you kind of spending money 615 00:33:34,040 --> 00:33:36,000 with your credit card? You have your smartphone that's 616 00:33:36,000 --> 00:33:39,840 with your 24/7, but there's 50 million cameras in the US. 617 00:33:39,840 --> 00:33:42,400 They kept you pretty much everywhere that you go there's 618 00:33:42,400 --> 00:33:44,120 unless you live. Out there in the. 619 00:33:44,120 --> 00:33:46,400 Woods and grow your own vegetables. 620 00:33:46,680 --> 00:33:49,720 There's just no way that you can escape it, and I think Minority 621 00:33:49,880 --> 00:33:52,320 Report in a way shows. This right, it's like not that 622 00:33:52,960 --> 00:33:54,600 he kind of. Post something on social media 623 00:33:54,600 --> 00:33:56,240 and someone infers that you're extroverted. 624 00:33:56,440 --> 00:33:58,960 Yeah, he just walks. By you get like your record 625 00:33:58,960 --> 00:34:01,280 faces recognized. He connects it to everything 626 00:34:01,440 --> 00:34:03,840 that they know about you just by looking at the data traces that 627 00:34:03,840 --> 00:34:06,520 you leave unintentionally. And I think that's the world 628 00:34:06,520 --> 00:34:09,080 that we live in today. So that's why I like like this 629 00:34:09,159 --> 00:34:11,040 example in particular and the Tik. 630 00:34:11,040 --> 00:34:11,880 T.O.K. Example. 631 00:34:12,000 --> 00:34:15,719 There's not as much research on the assumption that algorithms 632 00:34:15,719 --> 00:34:17,960 can know you better than you know yourself. 633 00:34:18,440 --> 00:34:21,159 But the interesting part here is I think that algorithms just see 634 00:34:21,159 --> 00:34:22,679 a lot more, right? So what did? 635 00:34:22,679 --> 00:34:25,159 The power of AI comes from the fact that they observe your 636 00:34:25,159 --> 00:34:27,480 behavior, but they also observe the behavior of. 637 00:34:27,480 --> 00:34:30,560 Millions of people so there's certain patterns that they 638 00:34:30,560 --> 00:34:33,719 couldn't detect is like these digital Sherlock Holmes's and 639 00:34:33,719 --> 00:34:36,280 that you might just not have access to and because there's 640 00:34:36,280 --> 00:34:39,239 all these social pressures to. Conform in a certain way. 641 00:34:39,679 --> 00:34:42,120 You just might not be top of mind, but an algorithm just 642 00:34:42,120 --> 00:34:45,400 observes everybody and now you can see OK based on these kind 643 00:34:45,400 --> 00:34:48,000 of more subtle cues. Maybe that's something that I 644 00:34:48,000 --> 00:34:51,360 recommend and maybe that makes you wake up to a reality that 645 00:34:51,360 --> 00:34:53,440 you didn't even see yourself. So I think it's an interesting 646 00:34:53,440 --> 00:34:56,920 one that we don't fully understand yet, but I think 647 00:34:56,920 --> 00:35:01,760 we're getting there. If you'll allow me, I want to do 648 00:35:01,760 --> 00:35:04,160 a quick sidebar on Cambridge Analytica. 649 00:35:04,320 --> 00:35:07,120 We actually, we talked about this in our episode with Gordon 650 00:35:07,120 --> 00:35:09,680 Pennycook when we were talking about misinformation. 651 00:35:09,680 --> 00:35:13,560 So totally different context. And can you help us set the 652 00:35:13,560 --> 00:35:15,920 record straight? This also applies to a general, 653 00:35:16,280 --> 00:35:17,920 you know, question we have in terms of impact. 654 00:35:17,920 --> 00:35:20,320 What is the impact of psychological targeting in 655 00:35:20,320 --> 00:35:22,440 general? You know, some in some cases 656 00:35:22,480 --> 00:35:25,160 quite powerful, in other cases maybe less. 657 00:35:25,160 --> 00:35:30,160 So for the individual, what is your take on how much of an 658 00:35:30,160 --> 00:35:34,200 impact the targeting that Cambridge Analytica did? 659 00:35:34,840 --> 00:35:36,600 How much did that contribute? To. 660 00:35:37,000 --> 00:35:38,760 Donald Trump being. Elected. 661 00:35:39,120 --> 00:35:41,320 I mean, there's like a few thoughts that I have on on 662 00:35:41,360 --> 00:35:43,640 Cambridge Analytica. The first one, I mean, it's not 663 00:35:43,640 --> 00:35:47,000 impossible to say what actual impact it had on the election. 664 00:35:47,160 --> 00:35:48,760 I think for me, that was never the question. 665 00:35:48,840 --> 00:35:51,560 I was actually really annoyed with the news coverage of like, 666 00:35:51,560 --> 00:35:53,120 well, did they do it? Did they not do it? 667 00:35:53,120 --> 00:35:55,520 I'm like, I don't care. I know that the technology 668 00:35:55,520 --> 00:35:57,480 exists. So even if they didn't do it the 669 00:35:57,480 --> 00:36:00,280 next around, someone is going to use it, right? 670 00:36:00,280 --> 00:36:02,280 And if we believe, and I think this is getting to your 671 00:36:02,280 --> 00:36:06,040 question, do I think that again, we can convince a die hard 672 00:36:06,040 --> 00:36:08,160 Hillary supporter to suddenly vote for Trump? 673 00:36:08,480 --> 00:36:11,560 It's probably not, right? This is just we're not changing 674 00:36:11,560 --> 00:36:14,120 people's core identity. We're not kind of flipping you 675 00:36:14,120 --> 00:36:17,000 around and turn you into a person that's someone totally 676 00:36:17,000 --> 00:36:19,600 different just by showing you a few personalized ads on 677 00:36:19,600 --> 00:36:21,520 Facebook. That's not how the world works. 678 00:36:22,000 --> 00:36:23,480 But most of the time we don't have to do that. 679 00:36:23,480 --> 00:36:26,880 But most of the people are either undecided on who to vote 680 00:36:26,880 --> 00:36:29,440 for or they're kind of considering whether to go out on 681 00:36:29,440 --> 00:36:31,920 Election Day at all to, to cast their votes. 682 00:36:31,920 --> 00:36:35,120 So I think in that election is like 40% of people didn't even 683 00:36:35,120 --> 00:36:37,680 bother to go to the ballots and then cast their votes. 684 00:36:37,680 --> 00:36:42,640 So can we probably nudge them to either change their opinions on 685 00:36:42,640 --> 00:36:47,520 the candidates or decide to stay at home, which is what what the 686 00:36:48,280 --> 00:36:50,760 Cambridge Analytica campaign was trying to do? 687 00:36:51,640 --> 00:36:53,640 That's probably true. And if you look at the 688 00:36:53,640 --> 00:36:56,240 elections, most of them are won by really small margins. 689 00:36:56,520 --> 00:36:59,680 So I don't think it's like this magical warfare tool that it was 690 00:36:59,680 --> 00:37:01,680 made out to be. But I do think that it could 691 00:37:01,680 --> 00:37:04,160 potentially have an impact even on something as big as 692 00:37:04,160 --> 00:37:07,400 elections, just because that the margins of error of which side 693 00:37:07,400 --> 00:37:09,880 wins are just so, so small. So that's one thing. 694 00:37:10,000 --> 00:37:12,760 The other one that I think is really important in that the 695 00:37:12,760 --> 00:37:17,040 coverage was it was always made out to be like this data leak. 696 00:37:17,040 --> 00:37:19,720 So I think that the, the public narrative was like, Oh my God, 697 00:37:19,720 --> 00:37:22,760 someone hacked into Facebook, they got all of your data and 698 00:37:22,760 --> 00:37:26,400 that was like this tragedy that wasn't true, that the way that 699 00:37:26,400 --> 00:37:28,080 the data was acquired was perfect. 700 00:37:28,080 --> 00:37:30,520 Well, partially, and I'm going to explain in a second why it 701 00:37:30,520 --> 00:37:32,840 wasn't. But back in the day, because 702 00:37:32,840 --> 00:37:35,960 Facebook was trying to connect everybody, like the moment that 703 00:37:35,960 --> 00:37:39,680 I consented to have you access my profile, you also through the 704 00:37:39,680 --> 00:37:42,800 Graph API got access to the profiles of all of my friends. 705 00:37:43,240 --> 00:37:46,760 So Cambridge Analytica didn't have to kind of ask every single 706 00:37:46,760 --> 00:37:49,080 person legally to get their Facebook profile. 707 00:37:49,080 --> 00:37:52,040 They could access that through my consent and through my 708 00:37:52,040 --> 00:37:54,320 profile. Now, the problem with Cambridge 709 00:37:54,320 --> 00:37:57,240 Analytica was that they collected the data for different 710 00:37:57,240 --> 00:37:59,960 purpose, right? It was an academic collecting it 711 00:37:59,960 --> 00:38:02,000 and then selling it on to Cambridge Analytica. 712 00:38:02,000 --> 00:38:05,240 So that was the illegal part. But the initial way that the 713 00:38:05,240 --> 00:38:07,920 data was collected was actually in line with the terms and 714 00:38:07,920 --> 00:38:09,720 conditions of Facebook. And I think that's something 715 00:38:09,720 --> 00:38:13,160 that we need to understand. It often times the way that our 716 00:38:13,160 --> 00:38:16,920 data is being captured isn't through a security breach or a 717 00:38:16,920 --> 00:38:18,560 data leak, right? Often times we think, oh, 718 00:38:18,560 --> 00:38:21,160 someone is going to hack into our e-mail or it's going to get 719 00:38:21,160 --> 00:38:22,920 our data. Most of the time we just sign 720 00:38:22,920 --> 00:38:26,240 away our data by accepting the terms and conditions so that you 721 00:38:26,400 --> 00:38:29,640 download a weather app on your phone and you happily accept 722 00:38:29,920 --> 00:38:32,080 what the app to tap into your microphone cause. 723 00:38:32,120 --> 00:38:33,920 The gallery, the terms and conditions. 724 00:38:33,960 --> 00:38:36,600 Nobody's reading it exactly like nobody has the time. 725 00:38:37,080 --> 00:38:41,160 It feels like a leak because you didn't know that was possible, 726 00:38:41,520 --> 00:38:45,520 but it wasn't actually. It's just poor design of privacy 727 00:38:45,520 --> 00:38:47,120 and security in the 1st place, right? 728 00:38:47,240 --> 00:38:49,240 Exactly. And I think that makes a big 729 00:38:49,240 --> 00:38:53,600 difference. Yeah, cuz you can't really go 730 00:38:53,600 --> 00:38:56,880 out now legally and prosecute those companies because 731 00:38:56,880 --> 00:38:59,080 essentially you consented to it. So I think maybe we're going to 732 00:38:59,080 --> 00:39:01,600 get into this later, but I think we just have to design the 733 00:39:01,920 --> 00:39:04,760 redesign the entire process right now we're just saying, 734 00:39:04,760 --> 00:39:07,560 well, we're going to tell you a little bit about about which 735 00:39:07,560 --> 00:39:09,480 data we're collecting, what we're doing with it, and then 736 00:39:09,480 --> 00:39:12,640 you take control over how to manage it. 737 00:39:12,840 --> 00:39:14,200 It's just impossible. Right. 738 00:39:14,200 --> 00:39:16,640 So even if you understand exactly what's happening, what 739 00:39:16,640 --> 00:39:20,720 data they're taking, you don't have the time to manage your own 740 00:39:20,720 --> 00:39:23,920 data across all of the services and products that you're using. 741 00:39:23,920 --> 00:39:27,960 It's just impossible. If there's no default level of 742 00:39:28,000 --> 00:39:33,160 of privacy protection that is is kind of at a good level, it's 743 00:39:33,160 --> 00:39:36,520 just not going to happen. Yeah. 744 00:39:36,560 --> 00:39:38,360 OK, Tangent over on Cambridge Analytica. 745 00:39:38,360 --> 00:39:40,480 Yeah. But a very good tangent and I 746 00:39:40,480 --> 00:39:44,240 think, you know, I'm obviously we're super tempted to jump into 747 00:39:44,240 --> 00:39:46,480 talking about AI. This isn't. 748 00:39:46,760 --> 00:39:49,000 But actually I want to continue to lay a little bit of 749 00:39:49,000 --> 00:39:52,000 foundation because I think it's interesting given that we're now 750 00:39:52,000 --> 00:39:55,240 introduced kind of what you mentioned before about it's kind 751 00:39:55,240 --> 00:39:58,720 of like call it 2 main steps. It would be interesting to look 752 00:39:58,720 --> 00:40:01,320 at that a little more in detail because I think especially when 753 00:40:01,320 --> 00:40:06,080 it comes to, you know, citing another maybe example that many 754 00:40:06,080 --> 00:40:08,720 people have seen is this film that came out maybe two years 755 00:40:08,720 --> 00:40:11,040 ago, Social Dilemma was it out maybe two years ago? 756 00:40:11,920 --> 00:40:15,360 And they have this very strong visual kind of metaphor where 757 00:40:15,360 --> 00:40:18,600 they're showing this main. Protagonist of sort and you have 758 00:40:18,600 --> 00:40:21,960 this one with social media company that like through his 759 00:40:22,160 --> 00:40:26,320 interacting in the world, it'd be like a more and more of a one 760 00:40:26,320 --> 00:40:29,200 to one kind of representation of him in their system. 761 00:40:29,200 --> 00:40:33,040 So initially it's like very kind of not really detailed and as 762 00:40:33,040 --> 00:40:35,640 the film progresses, like it becomes a very kind of 1 to one 763 00:40:35,640 --> 00:40:38,920 representation of of him, but it's still kind of interesting 764 00:40:38,920 --> 00:40:40,880 in terms of like, well, what does that actually mean? 765 00:40:40,880 --> 00:40:43,200 Like what if we wanted to build representations of people? 766 00:40:43,640 --> 00:40:45,400 What are the things that we might look at? 767 00:40:45,400 --> 00:40:47,960 And I think you cited the idea of simple, you know, 768 00:40:48,000 --> 00:40:50,200 extroversion versus introversion, which is a kind of 769 00:40:50,200 --> 00:40:52,520 a some form of personality trait. 770 00:40:52,960 --> 00:40:56,000 And I guess it'll be interesting to talk a little bit more about 771 00:40:56,000 --> 00:40:57,440 that. You know, you could imagine that 772 00:40:57,880 --> 00:40:59,800 the type of things that could be included could be everything 773 00:40:59,800 --> 00:41:04,520 from I guess people's kind of preferences, maybe both kind of 774 00:41:05,080 --> 00:41:08,880 reveal and stated it could be as we're talking about personality, 775 00:41:09,240 --> 00:41:11,240 maybe some other things. I just want to see what would 776 00:41:11,240 --> 00:41:14,720 you say is kind of the main thing that is being used at the 777 00:41:14,720 --> 00:41:17,120 kind of the first stage of like creating that picture? 778 00:41:17,120 --> 00:41:20,240 Like what are the main things that usually kind of part of 779 00:41:20,240 --> 00:41:21,840 that equation? That's a great. 780 00:41:21,840 --> 00:41:24,680 Question and and the extroversion versus introversion 781 00:41:24,680 --> 00:41:26,960 was a gross oversimplification, I should say. 782 00:41:27,280 --> 00:41:29,200 So I'm. A personality psychologist. 783 00:41:29,200 --> 00:41:31,640 And and psychometrician by trainings that would have get me 784 00:41:31,640 --> 00:41:33,040 kicked out of the Society of that. 785 00:41:33,360 --> 00:41:36,440 But I think there's a number of different steps. 786 00:41:36,440 --> 00:41:40,200 So one is, I think like any social scientist or behavioral 787 00:41:40,200 --> 00:41:43,000 scientist would say, well, why don't you just go from behavior, 788 00:41:43,280 --> 00:41:45,160 past behavior to future behavior, right? 789 00:41:45,160 --> 00:41:47,920 If you can observe someone's entire behavioral history, why 790 00:41:47,920 --> 00:41:50,800 don't you just use that to make predictions about what they're 791 00:41:50,800 --> 00:41:53,200 going to do next? And that's totally true, right? 792 00:41:53,200 --> 00:41:56,440 I think we don't necessarily always need the intermediary 793 00:41:56,440 --> 00:41:58,760 step of saying we were translating your behavioral 794 00:41:58,760 --> 00:42:02,200 profile into something like the Big 5 personality traits, your 795 00:42:02,200 --> 00:42:04,680 moral foundations, your values, whatever it is. 796 00:42:05,160 --> 00:42:07,640 Now, I do think that's actually sometimes it's helpful because 797 00:42:07,640 --> 00:42:10,120 it gives us a human understanding, right? 798 00:42:10,120 --> 00:42:12,720 And a lot of the interactions that we have, especially in the 799 00:42:12,720 --> 00:42:15,480 second step when we're trying to influence behavior is still 800 00:42:15,480 --> 00:42:19,800 driven by humans. So if I'm a marketing creative, 801 00:42:19,840 --> 00:42:24,240 I might take the insights into who someone is on again, 802 00:42:24,240 --> 00:42:26,560 personality level to create some messages. 803 00:42:26,560 --> 00:42:29,000 Now we're going to talk about AI in a second. 804 00:42:29,080 --> 00:42:32,360 So some of this is actually being automated now by by these 805 00:42:32,360 --> 00:42:35,080 large language models, but there's still something that is 806 00:42:35,160 --> 00:42:37,920 really meaningful, especially if we want to bridge this kind of 807 00:42:37,920 --> 00:42:40,520 experience between offline and online world to have an 808 00:42:40,520 --> 00:42:42,400 understanding that humans can make sense of. 809 00:42:42,920 --> 00:42:45,720 Now, I think you mentioned personality traits, moral 810 00:42:45,720 --> 00:42:48,560 foundations, preferences. So the extent to which these 811 00:42:48,560 --> 00:42:52,040 psychological traits are relevant and helpful when we 812 00:42:52,080 --> 00:42:55,440 think about catering content actually very much depends on 813 00:42:55,440 --> 00:42:57,400 the context, right? So the big 5 are often times 814 00:42:57,400 --> 00:43:00,600 helpful because they're so broad, like anything from music 815 00:43:00,600 --> 00:43:02,640 preferences to vocational interest. 816 00:43:02,640 --> 00:43:04,560 They kind of explain some of the variance. 817 00:43:05,000 --> 00:43:06,880 They're also really crude, right? 818 00:43:06,880 --> 00:43:09,040 Like if I think of like personalities, like you meet a 819 00:43:09,040 --> 00:43:11,200 person in the street, you know nothing about them. 820 00:43:11,480 --> 00:43:13,800 Knowing that they're extroverted is actually helpful because it 821 00:43:13,800 --> 00:43:15,560 gives you at least a sense of who they are. 822 00:43:15,720 --> 00:43:18,040 The moment that you get to know them a lot better, you throw the 823 00:43:18,040 --> 00:43:21,280 big 5 out the window because you don't think about your spouse is 824 00:43:21,280 --> 00:43:24,880 like, Oh my God, he's so extra. You know them so much better and 825 00:43:24,880 --> 00:43:27,240 on every little detail that you don't need this anymore. 826 00:43:27,720 --> 00:43:30,160 So I think this is one thing, but you can also imagine that 827 00:43:30,160 --> 00:43:34,280 other contexts, for example, in think about health, I don't 828 00:43:34,280 --> 00:43:36,720 think the big 5 are necessarily the main driver you could think 829 00:43:36,720 --> 00:43:39,320 of, well, if I wanna change behaviors, maybe I'm gonna go 830 00:43:39,320 --> 00:43:42,480 for regulatory focus. Are people more motivated by 831 00:43:42,920 --> 00:43:46,320 prevention or by promotion? So I think this very much 832 00:43:46,320 --> 00:43:48,640 depends what you're trying to measure on what you're trying to 833 00:43:48,640 --> 00:43:50,440 accomplish in terms of behavior change. 834 00:43:51,240 --> 00:43:54,520 But even that I still think assumes that we like the static 835 00:43:54,520 --> 00:43:57,520 version of ourselves, right? I think a lot of personality 836 00:43:57,760 --> 00:44:01,000 psychology initially was let's just capture someone's 837 00:44:01,000 --> 00:44:03,400 dispositions. Let's just see what are the 838 00:44:03,400 --> 00:44:06,680 stable dimensions that underline our behavior, thoughts and and 839 00:44:06,680 --> 00:44:08,240 feelings. And I think personality 840 00:44:08,240 --> 00:44:11,600 psychology has shifted quite a bit over the last years to an 841 00:44:11,600 --> 00:44:14,080 understanding that I think is first of all, a lot more 842 00:44:14,080 --> 00:44:17,520 accurate and a lot more nuanced. So this idea that we're not 843 00:44:17,520 --> 00:44:19,600 always the same, right? This is always the debate going 844 00:44:19,600 --> 00:44:21,760 on between personal psychologists and social 845 00:44:21,760 --> 00:44:25,080 psychologist, where personality psychologist, it's the person 846 00:44:25,080 --> 00:44:28,320 and social psychologist was like, no, it's the context that 847 00:44:28,320 --> 00:44:30,920 determines how we behave. Could it be both? 848 00:44:31,080 --> 00:44:34,600 Exactly. And I think that's a model that 849 00:44:34,600 --> 00:44:36,760 we have now personality that I really like. 850 00:44:36,760 --> 00:44:40,120 So the way that we think about it today is it's not just a one 851 00:44:40,120 --> 00:44:42,640 time measure. It's essentially if I sampled 852 00:44:42,640 --> 00:44:45,840 how extroverted you feel behave across many different points in 853 00:44:45,840 --> 00:44:48,920 time, what I would get is a distribution, right? 854 00:44:48,920 --> 00:44:51,640 So you have like a distribution that's looks relatively normally 855 00:44:51,640 --> 00:44:53,320 distributed. You have a lot of stuff that's 856 00:44:53,320 --> 00:44:55,800 centered around a certain mean. So if you're more extroverted, 857 00:44:55,800 --> 00:44:57,840 your mean is higher. If you're more introverted, your 858 00:44:57,840 --> 00:45:01,000 mean is lower. But it still means that we give 859 00:45:01,000 --> 00:45:04,360 you at least a chance to be slightly more extroverted than 860 00:45:04,360 --> 00:45:06,400 you typically are. Maybe when you're in a social 861 00:45:06,400 --> 00:45:08,440 context, right? You're at a party, there's other 862 00:45:08,440 --> 00:45:10,400 people, maybe you feel more extroverted than usual. 863 00:45:10,760 --> 00:45:14,040 If you're at the library or at home, you probably feel less 864 00:45:14,040 --> 00:45:17,000 extroverted because there's just not as much social stimulation. 865 00:45:17,240 --> 00:45:21,520 So I think what this means is we can get a much more one's 866 00:45:21,520 --> 00:45:24,160 understanding of a person in a given point in time. 867 00:45:24,560 --> 00:45:27,280 So I understand that most of the time you're relatively 868 00:45:27,280 --> 00:45:29,680 extroverted. But right now I can adjust my 869 00:45:29,680 --> 00:45:32,520 prediction of who you are because I know that you're not 870 00:45:32,520 --> 00:45:34,920 in a social setting. And this is true for humans, 871 00:45:34,920 --> 00:45:36,000 right? So we can make these 872 00:45:36,000 --> 00:45:38,000 adjustments. The same is true for algorithms, 873 00:45:38,680 --> 00:45:42,240 like if an algorithm gets hold of your time behavior history, 874 00:45:42,240 --> 00:45:44,840 they might get a good sense of your general disposition to be 875 00:45:44,840 --> 00:45:47,360 extroverted. Now, if they can combine this 876 00:45:47,520 --> 00:45:50,000 with knowledge of some of the momentary data that you 877 00:45:50,000 --> 00:45:52,200 generate, right, it can tap into your GPS. 878 00:45:52,360 --> 00:45:54,840 You can figure out that right now you're at a certain 879 00:45:54,840 --> 00:45:59,480 location, it's a bar, it's usually buzzing full at 5:00 PM 880 00:45:59,480 --> 00:46:01,800 and that's where you are. Maybe it's going to adjust the 881 00:46:01,800 --> 00:46:04,360 prediction to make you a little bit more extroverted. 882 00:46:04,440 --> 00:46:06,640 And who knows, Maybe that's the time when I should be targeting 883 00:46:06,640 --> 00:46:09,920 you with like an extroverted product because you're self 884 00:46:10,080 --> 00:46:11,640 concept of extroversion is activated. 885 00:46:11,920 --> 00:46:13,760 But it could also be that I should actually wait when you 886 00:46:13,760 --> 00:46:16,560 feel a little bit less extroverted than usual, maybe 887 00:46:16,560 --> 00:46:19,680 that's when you're craving more extroversion and that's when you 888 00:46:19,680 --> 00:46:22,320 kind of buy more likely to buy. So I think this dynamic 889 00:46:22,320 --> 00:46:25,560 understanding of of personality is where we're headed. 890 00:46:25,600 --> 00:46:27,360 That's cool. And this allows me to do a 891 00:46:27,360 --> 00:46:30,920 little bit of a kind of my turn to do a little bit of a detour 892 00:46:30,920 --> 00:46:35,200 or like tangent, no, but because you mentioned distribution here. 893 00:46:35,200 --> 00:46:36,800 And I think that's super interesting when it comes to 894 00:46:36,800 --> 00:46:40,360 personality, because I think a lot of people, if they're not 895 00:46:40,360 --> 00:46:42,920 like you or if they're not as experienced with personality, 896 00:46:42,920 --> 00:46:46,240 they might be exposed to some of the less. 897 00:46:46,880 --> 00:46:48,760 How do you say good models? But more popular. 898 00:46:49,440 --> 00:46:54,480 Less robust and where people are placed in kind of very rigid. 899 00:46:55,680 --> 00:46:58,160 Dichotomies. Let's name it. 900 00:46:58,160 --> 00:46:59,920 It's the MBTI. Yeah, what we're talking about. 901 00:46:59,920 --> 00:47:03,120 Well, that's one of them. I'm in Sweden and so in Sweden I 902 00:47:03,120 --> 00:47:06,520 have this hate relationship with a book that is best selling book 903 00:47:06,520 --> 00:47:08,600 for the last 10 years, which is about the disk model. 904 00:47:09,000 --> 00:47:11,840 So it's even worse the disk honestly dynamic. 905 00:47:11,840 --> 00:47:14,760 Yeah. So but whether they're placing a 906 00:47:14,760 --> 00:47:18,800 collar or as a letter, they're placed in that specifically. 907 00:47:18,800 --> 00:47:23,000 But obviously with Big Fiber Ocean, the whole interesting 908 00:47:23,000 --> 00:47:25,760 part is that we're placing a continuum of sort or 909 00:47:25,760 --> 00:47:29,680 distribution rather. And you know, looking at my own, 910 00:47:29,680 --> 00:47:32,600 I think it's interesting where there's 2 dimensions where I'm 911 00:47:32,600 --> 00:47:37,280 like at really at the kind of like high percentile, but some 912 00:47:37,280 --> 00:47:39,880 of them I'm kind of close to the middle. 913 00:47:40,400 --> 00:47:44,560 And I guess technically I would probably be categorized as an 914 00:47:44,560 --> 00:47:45,440 extrovert. You're. 915 00:47:45,440 --> 00:47:49,000 Definitely an extrovert, Sam. I score maybe 55. 916 00:47:49,120 --> 00:47:50,080 I score like 60. Ludicrous. 917 00:47:52,400 --> 00:47:54,680 Yeah. Definitely an extrovert. 918 00:47:55,640 --> 00:47:58,520 I don't know about that. But then like on something else 919 00:47:58,520 --> 00:48:01,080 like conscientious, I'm like at 9095 or something. 920 00:48:01,080 --> 00:48:04,480 So like technically if you have to bookmark me, I would be an 921 00:48:04,480 --> 00:48:06,160 extrovert and conscientious, but. 922 00:48:06,280 --> 00:48:08,160 Yeah. They're very different in terms 923 00:48:08,160 --> 00:48:09,920 of where I am on the distribution. 924 00:48:09,960 --> 00:48:12,440 And so that's my question basically like how does that fit 925 00:48:12,440 --> 00:48:13,400 into this? Like is that? 926 00:48:13,560 --> 00:48:16,160 Yeah, yeah. It's such a great question cuz 927 00:48:16,360 --> 00:48:19,000 technically, right. So if you start with the data, 928 00:48:19,280 --> 00:48:22,600 we're like million dimensional. So every single data point is in 929 00:48:22,600 --> 00:48:24,200 a way a dimension. We're already reducing 930 00:48:24,200 --> 00:48:27,440 complexity into five if we kind of take the big 5 again, which 931 00:48:27,520 --> 00:48:30,120 already reducing complexity into 5 dimensions. 932 00:48:30,240 --> 00:48:33,880 So that's a huge kind of step into kind of limiting the the 933 00:48:33,880 --> 00:48:36,520 amount of information that I can can see about you or can use 934 00:48:36,520 --> 00:48:39,000 about you. And then as you said with like 935 00:48:39,000 --> 00:48:41,920 some of the the questionnaires that have intuitive appeal, 936 00:48:41,920 --> 00:48:43,680 right? People love love labels. 937 00:48:43,880 --> 00:48:45,320 People love. Something that they can stick 938 00:48:45,320 --> 00:48:48,280 to, they can use it to explain. You can identify with that in 939 00:48:49,000 --> 00:48:51,160 case with the MBTI, but the problem is that you're 940 00:48:51,160 --> 00:48:54,120 dichotomizing something that shouldn't be dichotomized. 941 00:48:54,120 --> 00:48:55,560 And most of us are somewhere in the middle, right? 942 00:48:55,560 --> 00:48:57,800 There's very few people are extremely extroverted. 943 00:48:58,160 --> 00:49:00,040 Very few people are extremely introverted. 944 00:49:00,040 --> 00:49:02,200 So most of us. Hovering somewhere around the 945 00:49:02,200 --> 00:49:04,760 middle of the distribution, if that's the cut off, we're just 946 00:49:04,760 --> 00:49:07,880 gonna make a lot of mistakes. I think in the context of 947 00:49:07,880 --> 00:49:11,560 personality, what's interesting, and I don't think we we actually 948 00:49:11,560 --> 00:49:15,880 know the answer yet is how do we best personalize, right? 949 00:49:15,880 --> 00:49:19,480 Again, like if I were to talk to my marketing colleagues, I think 950 00:49:19,480 --> 00:49:22,200 the typical argument, the days, the more information you have 951 00:49:22,400 --> 00:49:25,600 the better. So if that's true, I would be 952 00:49:25,600 --> 00:49:28,480 much better off taking your entire big 5 personality profile 953 00:49:28,480 --> 00:49:31,320 because it tells me something about your extra version, your 954 00:49:31,320 --> 00:49:32,480 conscientiousness and everything. 955 00:49:32,960 --> 00:49:37,000 However, it could also be true that the ratio of signal to 956 00:49:37,000 --> 00:49:39,680 noise, it just becomes much, much higher. 957 00:49:39,680 --> 00:49:42,000 So lower, so cloudy. Exactly. 958 00:49:42,000 --> 00:49:44,000 It becomes too cloudy because like what does it tell me? 959 00:49:44,000 --> 00:49:47,440 If you're like a six sixty 60th percentile on on extra version, 960 00:49:47,600 --> 00:49:50,880 not entirely show what you make of and it might be better to 961 00:49:50,880 --> 00:49:53,280 just go for the ones that are extremely salient, right, 962 00:49:53,280 --> 00:49:55,320 because the the the ones that are extremely salient. 963 00:49:55,320 --> 00:49:58,680 So you're 95 percentile of conscientiousness. 964 00:49:59,000 --> 00:50:02,080 That's probably something that's driving your behavior quite 965 00:50:02,080 --> 00:50:04,560 dramatically, right? Because it's a core of who you 966 00:50:04,560 --> 00:50:05,760 are. Otherwise you wouldn't have take 967 00:50:05,760 --> 00:50:08,280 strongly agree on pretty much all of the questions. 968 00:50:08,520 --> 00:50:10,960 And so we might actually be better off ignoring some of the 969 00:50:10,960 --> 00:50:13,800 trades in the middle and just focus on the ones that seem to 970 00:50:13,800 --> 00:50:15,400 be driving your behavior the most. 971 00:50:15,680 --> 00:50:17,080 And again, we don't really know yet. 972 00:50:17,160 --> 00:50:19,160 I think it's an interesting question that we haven't looked 973 00:50:19,160 --> 00:50:22,680 into empirically. My hunch is that it's one of 974 00:50:22,680 --> 00:50:25,720 these instances where discarding some of the information could 975 00:50:25,720 --> 00:50:27,680 actually be could actually make you better off. 976 00:50:29,120 --> 00:50:32,200 Super interesting. Thanks for diving even deeper 977 00:50:32,200 --> 00:50:32,920 there. That's. 978 00:50:32,920 --> 00:50:34,400 Awesome. And we might get back to this in 979 00:50:34,400 --> 00:50:37,840 the context of AI, cuz the one thing that we know is like from 980 00:50:37,840 --> 00:50:43,080 some of my recent work is like language models like ChatGPT are 981 00:50:43,080 --> 00:50:46,080 extremely good, like crafting messages. 982 00:50:46,080 --> 00:50:49,800 If I give them, if I tell them, look, Samuel, he's extroverted. 983 00:50:49,800 --> 00:50:52,360 Like just come up with an iPhone ad that's tailored to his 984 00:50:52,360 --> 00:50:54,600 extroverted personality. Extremely good. 985 00:50:54,600 --> 00:50:55,920 But I don't even need to tell it. 986 00:50:56,120 --> 00:50:57,600 Here's what extroversion is all about. 987 00:50:57,680 --> 00:50:59,720 It's obviously right. The entire Internet, so it 988 00:50:59,720 --> 00:51:00,720 knows. What? 989 00:51:00,720 --> 00:51:03,360 Extroversion is all about, and it's great at crafting these 990 00:51:03,360 --> 00:51:05,920 messages. Now one thing that we tried 991 00:51:05,920 --> 00:51:08,560 initially was to say, OK, let's add more new ones because now 992 00:51:08,560 --> 00:51:11,680 that we are not relying on human creative people, maybe we can 993 00:51:11,680 --> 00:51:14,080 just take the full personality profile and we could just do 994 00:51:14,080 --> 00:51:17,120 this for every person, right? So I can create a very specific 995 00:51:17,120 --> 00:51:20,080 ad that's just targeted at you and like in your 5 dimensional 996 00:51:20,080 --> 00:51:22,920 personality space. Now what happened is that 997 00:51:22,920 --> 00:51:25,920 essentially the way the Cha Chi PT did it, it just kind of took 998 00:51:25,920 --> 00:51:28,680 every single trade and had like one sentence for every single 999 00:51:28,680 --> 00:51:31,160 trade. So instead of integrating and 1000 00:51:31,160 --> 00:51:34,760 saying, well, OK, maybe an open minded extrovert is very 1001 00:51:34,760 --> 00:51:37,400 different to an open minded introvert, right? 1002 00:51:37,400 --> 00:51:39,880 The open minded introvert goes to the museum, open minded 1003 00:51:39,880 --> 00:51:42,960 extrovert is out there in the street with like a flamboyant 1004 00:51:42,960 --> 00:51:45,720 outfit. So they didn't really, it just 1005 00:51:45,720 --> 00:51:49,520 kind of additively said, OK, here's a sense fixed version, 1006 00:51:49,520 --> 00:51:54,840 here's a Center for openness. And in that case, it just, it 1007 00:51:54,840 --> 00:51:58,960 felt super awkward, right? It's not a very natural ad that 1008 00:51:58,960 --> 00:52:01,320 you see when it's just like 5 different statements. 1009 00:52:01,360 --> 00:52:03,000 Maybe the language models are going to get there at some 1010 00:52:03,000 --> 00:52:05,880 point, but at least right now you're better off just focusing 1011 00:52:05,880 --> 00:52:07,840 on one when you use these generative models. 1012 00:52:12,360 --> 00:52:16,160 So it's, it feels to me like we have a tendency to focus a lot 1013 00:52:16,160 --> 00:52:19,280 on the person of the personality and their values and so on. 1014 00:52:19,280 --> 00:52:22,360 But as you mentioned earlier, there's also a, a really 1015 00:52:22,360 --> 00:52:25,360 important part of the environment, the situation that 1016 00:52:25,360 --> 00:52:27,880 a person is in. And this can both influence 1017 00:52:27,880 --> 00:52:31,800 their personality and you know, what parts of them are amplified 1018 00:52:31,800 --> 00:52:34,880 and so on. But are there other aspects 1019 00:52:34,880 --> 00:52:37,680 that, that are influenced by the environment? 1020 00:52:37,680 --> 00:52:40,920 What can we sort of learn about behaviour? 1021 00:52:41,000 --> 00:52:44,520 What can we predict from something like GPS, someone's 1022 00:52:44,520 --> 00:52:47,880 location, or other sort of passive indicators of where 1023 00:52:47,880 --> 00:52:50,800 someone is, for example? I mean, it's super interesting 1024 00:52:50,800 --> 00:52:53,120 because first of all, I mean, GPS is just one of the data 1025 00:52:53,120 --> 00:52:55,960 points that I personally really like because it says so much 1026 00:52:55,960 --> 00:52:59,240 about who you are, right? So like GPS records, for 1027 00:52:59,240 --> 00:53:01,840 example, can tell you whether someone might be suffering from 1028 00:53:01,840 --> 00:53:04,680 depression because they're not leaving the house as much 1029 00:53:04,680 --> 00:53:06,120 anymore. It is much less physical 1030 00:53:06,120 --> 00:53:07,880 activity. So there's actually somewhere 1031 00:53:07,880 --> 00:53:11,880 that you can translate context into something that looks more 1032 00:53:11,880 --> 00:53:14,840 like a disposition or like a psychological dimension. 1033 00:53:15,200 --> 00:53:17,360 But also context is something that I think is really 1034 00:53:17,360 --> 00:53:22,240 interesting when we think about privacy because like one of the 1035 00:53:22,240 --> 00:53:24,240 big questions right now, right? If you look at all of the 1036 00:53:24,240 --> 00:53:28,000 companies like Facebook, Snapchat, Instagram, I think one 1037 00:53:28,000 --> 00:53:30,920 of the, the challenges that they have when it comes to privacy is 1038 00:53:30,920 --> 00:53:33,280 that they collect these vast amounts of data. 1039 00:53:33,600 --> 00:53:36,560 They collect essentially your entire history, everything that 1040 00:53:36,560 --> 00:53:40,200 you've done for years and years. And that data becomes incredibly 1041 00:53:40,200 --> 00:53:42,400 valuable, right? Because it's like a holistic 1042 00:53:42,400 --> 00:53:44,000 picture of everything that you've done. 1043 00:53:44,280 --> 00:53:47,200 So the moment that there's a data breach or security breach, 1044 00:53:47,600 --> 00:53:50,680 all of the data is gone. So one of the questions, I think 1045 00:53:51,000 --> 00:53:53,480 that's partially driven by regulation and the companies are 1046 00:53:53,480 --> 00:53:55,720 worried that at some point that it's just gonna stop, that 1047 00:53:55,720 --> 00:53:58,000 they're allowed to collect all of this behavioral data. 1048 00:53:58,600 --> 00:54:01,160 I think a lot of these companies are looking into alternatives 1049 00:54:01,160 --> 00:54:04,600 where you can fade out some of these behavioral histories by 1050 00:54:04,840 --> 00:54:07,880 considering context. So what I mean by that is 1051 00:54:08,200 --> 00:54:11,960 essentially, instead of me taking all of the data that 1052 00:54:11,960 --> 00:54:14,360 you've produced over the last two years and saying, well, 1053 00:54:14,360 --> 00:54:17,200 you're more extroverted or you're more introverted, maybe I 1054 00:54:17,200 --> 00:54:20,680 just skip most of that and say, OK, you're right now, forget 1055 00:54:20,680 --> 00:54:23,240 about who you are right now. You're in a context that should 1056 00:54:23,240 --> 00:54:27,080 make you relatively extroverted again, because you're at a bar 1057 00:54:27,080 --> 00:54:29,440 and there's, it's a social context and there's many people 1058 00:54:29,440 --> 00:54:31,840 and you're probably attuned to what other people think about 1059 00:54:31,840 --> 00:54:33,920 you and so on. So forget about who you are, 1060 00:54:34,000 --> 00:54:36,040 what the starting point is, we're just gonna take this 1061 00:54:36,040 --> 00:54:39,960 contextual information. And so with some of my students, 1062 00:54:39,960 --> 00:54:43,080 in collaboration with one of the bigger companies, we actually 1063 00:54:43,080 --> 00:54:44,680 tried just that. So we tried. 1064 00:54:45,000 --> 00:54:48,960 Can we equally well predict what someone is gonna do on the app 1065 00:54:48,960 --> 00:54:53,040 next by using these contextual cues and fading out some of 1066 00:54:53,040 --> 00:54:55,200 these more intimate personal histories? 1067 00:54:55,200 --> 00:54:57,920 And what you see is that it actually works pretty well. 1068 00:54:57,960 --> 00:55:02,760 So because a lot of the insights that I think you get about 1069 00:55:02,760 --> 00:55:05,080 people's psychology are captured by context, right? 1070 00:55:05,080 --> 00:55:08,480 So if I find you at a bar, well, that also means that your 1071 00:55:08,480 --> 00:55:11,560 extroverted disposition is probably a little bit higher 1072 00:55:11,560 --> 00:55:14,120 because you're at the bar and that was your place of choosing 1073 00:55:14,120 --> 00:55:16,560 most of the time. So a lot of the information that 1074 00:55:16,560 --> 00:55:19,200 we can get from context also tells us something about the 1075 00:55:19,200 --> 00:55:23,720 psychology without us having to collect and store for all 1076 00:55:23,720 --> 00:55:26,200 eternity all of the traces that you've ever created. 1077 00:55:26,200 --> 00:55:28,760 So I think it's an into context is like this very interesting 1078 00:55:29,240 --> 00:55:31,680 window into what's going on inside our mind in a current 1079 00:55:31,680 --> 00:55:34,800 moment that allows us to potentially protect user privacy 1080 00:55:34,800 --> 00:55:37,520 a little bit better while still being able to make predictions 1081 00:55:37,960 --> 00:55:40,320 about your behavior and the next steps that you might take. 1082 00:55:41,400 --> 00:55:43,560 That's true, though even introverts do get. 1083 00:55:43,640 --> 00:55:45,200 Once in a while, once in a while. 1084 00:55:45,800 --> 00:55:46,760 I know it is. Actually. 1085 00:55:46,760 --> 00:55:49,840 Some of my earliest research was looking at like the extent to 1086 00:55:49,840 --> 00:55:53,600 which aligning your spending with your psychology makes you 1087 00:55:53,600 --> 00:55:56,000 happier. And we created this experiment 1088 00:55:56,000 --> 00:55:59,480 where we forced introverts to go to the bar and have fun with 1089 00:55:59,480 --> 00:56:01,400 their friends, and it turned out that their happiness actually 1090 00:56:01,400 --> 00:56:02,600 dropped. After. 1091 00:56:02,600 --> 00:56:04,600 Doing so. So it's not only did it not make 1092 00:56:04,600 --> 00:56:07,000 them happier as it did for extroverts, it just actually 1093 00:56:07,000 --> 00:56:10,040 made them more miserable. So wow, Yeah, I I hear you. 1094 00:56:12,040 --> 00:56:16,200 OK, awesome. And in terms of those sort of 1095 00:56:16,200 --> 00:56:20,840 real time opportunities for interventions, can you think of 1096 00:56:20,840 --> 00:56:27,440 any other examples of like, you know, based on data from our the 1097 00:56:27,440 --> 00:56:30,520 sensors that are in our watches and our phones and those sorts 1098 00:56:30,520 --> 00:56:33,520 of things like if you're exercising or like, like what 1099 00:56:33,520 --> 00:56:37,120 other things can you be doing that we can learn about that can 1100 00:56:37,120 --> 00:56:41,520 maybe trigger or silence some kind of health intervention? 1101 00:56:42,280 --> 00:56:44,800 Yeah, for example, I think so for me, it's actually the 1102 00:56:44,800 --> 00:56:46,800 combination that's the most interesting part, right. 1103 00:56:46,800 --> 00:56:49,800 So if I can use your GPS records, for example, to say, 1104 00:56:50,160 --> 00:56:53,080 well, it looks like you might be entering a depressive episode 1105 00:56:53,080 --> 00:56:56,520 cuz again, you're deviating from your behavioral routine, you're 1106 00:56:56,520 --> 00:56:58,600 spending more time at home, you're not seeing friends as 1107 00:56:58,600 --> 00:57:01,280 much anymore. I have a better sense that you 1108 00:57:01,280 --> 00:57:04,120 might be struggling. Now the momentary aspect comes 1109 00:57:04,120 --> 00:57:07,640 in when I see you might be struggling and you're currently 1110 00:57:07,640 --> 00:57:11,440 walking close to a park. So now the best intervention is 1111 00:57:11,440 --> 00:57:13,640 just to tell you, like, look, seems like maybe you're not 1112 00:57:13,640 --> 00:57:15,800 having a great day. Why don't you just go and spend 1113 00:57:15,800 --> 00:57:18,200 some time in the park? And I think that's the 1114 00:57:18,200 --> 00:57:21,280 combination of like, what are the predictions about who we are 1115 00:57:21,280 --> 00:57:24,560 a little bit more generally? Plus, what's the context that 1116 00:57:24,560 --> 00:57:27,200 we're in right now that could drive interventions that just 1117 00:57:27,280 --> 00:57:29,800 are so much more convenient and so much more personal? 1118 00:57:30,360 --> 00:57:34,840 Yeah, much smarter. I'm so tempted to be like 1119 00:57:34,840 --> 00:57:38,160 putting you against the wall and being like, OK, if you would be 1120 00:57:38,160 --> 00:57:41,720 trying to get some person to, let's say, go to the gym and you 1121 00:57:41,720 --> 00:57:44,880 have to choose like either, you know, could lose equation. 1122 00:57:45,160 --> 00:57:49,520 You know, either you can take the the data you have on the P 1123 00:57:49,600 --> 00:57:52,400 the person, or you can only take the data you have on E the 1124 00:57:52,400 --> 00:57:55,080 environment. Which one would you choose if 1125 00:57:55,080 --> 00:57:57,520 you can only take one? I would probably, and for 1126 00:57:57,520 --> 00:58:00,440 different reasons, I would probably go with the context 1127 00:58:00,440 --> 00:58:03,120 because the reason for my context is so interesting in 1128 00:58:03,120 --> 00:58:06,600 this specific space is that it usually cues habits, right? 1129 00:58:06,600 --> 00:58:09,920 So the moment that I know you typically exercise, if this 1130 00:58:09,920 --> 00:58:13,280 happens when you walk past the gym, when you kind of get up 1131 00:58:13,280 --> 00:58:16,920 early in the morning, those are all cues that in a way get me to 1132 00:58:16,920 --> 00:58:18,600 a habit. So they kind of allow me to 1133 00:58:18,600 --> 00:58:22,320 trigger a habit that it's much harder to trigger with with like 1134 00:58:22,320 --> 00:58:25,200 the P component, with like the dispositional component. 1135 00:58:25,560 --> 00:58:28,320 So I think in the context of exercising, just because we know 1136 00:58:28,320 --> 00:58:31,280 that habits are so important and so critical in this case, I 1137 00:58:31,280 --> 00:58:33,000 would probably go with a contextual one. 1138 00:58:33,000 --> 00:58:34,640 What? What I want to do here is kind 1139 00:58:34,640 --> 00:58:37,560 of your thinking process around that, like how you think about 1140 00:58:37,560 --> 00:58:39,520 that. So that's really interesting to 1141 00:58:39,520 --> 00:58:43,120 hear. And yeah, I guess now should we? 1142 00:58:43,200 --> 00:58:47,800 Yeah, talk about AI. I mean, it's funny cuz I think 1143 00:58:47,800 --> 00:58:49,680 we've been talking about AI all the time, by the way. 1144 00:58:49,800 --> 00:58:53,880 So I think anything that's it's like machine learning is just a 1145 00:58:53,880 --> 00:58:57,680 subset of of AI. So, but I think there's like a, 1146 00:58:58,000 --> 00:59:01,840 yeah, a qualitatively different level of AI that we see now. 1147 00:59:02,000 --> 00:59:04,360 And I'm happy to get into that because I do think it's changing 1148 00:59:04,360 --> 00:59:06,480 this entire game of psychological targeting quite a 1149 00:59:06,480 --> 00:59:07,760 bit. Yeah. 1150 00:59:08,000 --> 00:59:10,920 Where would you start in terms of unpacking what has maybe 1151 00:59:11,400 --> 00:59:15,320 changed in some ways your world with the advancements of AI the 1152 00:59:15,320 --> 00:59:16,080 last few years? Yeah. 1153 00:59:16,680 --> 00:59:18,720 So I think we can actually go back all the way to the 1154 00:59:18,720 --> 00:59:22,240 beginning and the framework that we established, right? 1155 00:59:22,240 --> 00:59:25,560 If you think back those two steps, one is we take data and 1156 00:59:25,560 --> 00:59:26,920 make predictions about who you are. 1157 00:59:26,920 --> 00:59:29,600 And then the second step is based on these predictions, we 1158 00:59:29,600 --> 00:59:31,840 try and change your behavior and AI. 1159 00:59:31,840 --> 00:59:34,040 So now I'm talking mostly about generative AI. 1160 00:59:34,040 --> 00:59:39,240 So large language models like ChatGPT, they allow us to, in a 1161 00:59:39,240 --> 00:59:41,720 way, scale and democratize both parts. 1162 00:59:41,720 --> 00:59:44,720 So the first one, back in the day when I started, you 1163 00:59:44,720 --> 00:59:47,160 essentially had to have a proprietary model, right? 1164 00:59:47,160 --> 00:59:49,520 So if I wanted to get from your Facebook likes to your 1165 00:59:49,520 --> 00:59:53,400 personality, what I needed was, first of all, the data set that 1166 00:59:53,480 --> 00:59:56,680 got me access to your Facebook profile, your Facebook likes, 1167 00:59:56,680 --> 01:00:00,080 your status updates in combination with a self report 1168 01:00:00,360 --> 01:00:02,640 of your personality, right? So you also had to complete a 1169 01:00:02,640 --> 01:00:05,800 questionnaire to tell me, here's how I see myself and here's how 1170 01:00:05,800 --> 01:00:09,000 I would describe myself. And then I was able to start. 1171 01:00:09,120 --> 01:00:11,840 Training these models that say, OK, maybe if you like Lady Gaga, 1172 01:00:11,880 --> 01:00:14,440 you're more extroverted, if you like manga, you're more 1173 01:00:14,440 --> 01:00:16,720 introverted. But for that, I had to have the 1174 01:00:16,720 --> 01:00:18,640 data set and I had to train the model. 1175 01:00:19,240 --> 01:00:23,440 Now Fast forward to today, what you can do is you can 1176 01:00:23,440 --> 01:00:26,080 essentially just take someone's social media profile. 1177 01:00:26,120 --> 01:00:28,960 So just take all of those the status updates that someone, 1178 01:00:28,960 --> 01:00:31,480 someone post their Facebook likes, you can pop it into 1179 01:00:31,480 --> 01:00:34,680 ChatGPT and you can just ask it, OK, what do you, who do you 1180 01:00:34,680 --> 01:00:37,920 think a person who posted the stuff on Facebook and who 1181 01:00:37,920 --> 01:00:40,080 followed these pages, who do you think they are in terms of the 1182 01:00:40,080 --> 01:00:41,920 big 5? Just give me a numeric rating 1183 01:00:42,120 --> 01:00:45,040 from 1:00 to 7:00 on each of the five dimensions and it does a 1184 01:00:45,040 --> 01:00:47,360 remarkable job. So we just published this paper 1185 01:00:47,360 --> 01:00:50,640 showing that the accuracy by which ChatGPT, which was never 1186 01:00:50,640 --> 01:00:52,200 explicitly trained to do that, right? 1187 01:00:52,200 --> 01:00:55,920 So and nobody ever told ChatGPT get me from status updates to 1188 01:00:55,920 --> 01:00:59,600 personality profiles, but it still can do so with almost as 1189 01:00:59,600 --> 01:01:03,680 high in accuracy as if you train the model explicitly for this 1190 01:01:03,680 --> 01:01:06,000 purpose. And that is a total game 1191 01:01:06,000 --> 01:01:07,520 changer. So first of all, it means that 1192 01:01:07,520 --> 01:01:10,360 anyone can do it right. So back in the day, it was like 1193 01:01:10,360 --> 01:01:13,160 this outcry of well, Cambridge Analytica got access to all of 1194 01:01:13,160 --> 01:01:16,120 this data and Oh my God, they have these models like a 15 year 1195 01:01:16,120 --> 01:01:20,040 old kid in the parents basement can now scrape your data online, 1196 01:01:20,160 --> 01:01:22,880 run it through championship BT and essentially have like the 1197 01:01:23,000 --> 01:01:25,080 psychological profiles of millions of people. 1198 01:01:25,360 --> 01:01:28,000 And not just a big 5. But if you've, if we forgot to 1199 01:01:28,000 --> 01:01:31,960 collect moral foundations in our original data set, that's it. 1200 01:01:32,200 --> 01:01:34,600 We're not going to be able to predict it like chat. 1201 01:01:34,600 --> 01:01:36,800 GPD can predict whatever you wanted to predict. 1202 01:01:37,000 --> 01:01:39,760 You're not limited to the Big 5. You're not limited to what you 1203 01:01:39,760 --> 01:01:41,280 what data you collected initially. 1204 01:01:41,680 --> 01:01:45,120 It just knows everything about any psychological trade and you 1205 01:01:45,120 --> 01:01:46,880 can take it even a step further, right. 1206 01:01:46,880 --> 01:01:50,600 So back in the day, we just looked kind of passively and 1207 01:01:50,600 --> 01:01:52,920 looking backwards retrospectively at someone's 1208 01:01:52,920 --> 01:01:56,720 history to make predictions. Chat bots have the ability to 1209 01:01:56,720 --> 01:02:00,080 dynamically interact with you. So some of the more recent work 1210 01:02:00,320 --> 01:02:02,520 we actually looked at, well, can we train a chat bot to have a 1211 01:02:02,520 --> 01:02:05,440 conversation that still feels natural, like getting to know 1212 01:02:05,440 --> 01:02:08,360 someone on a first date, and then based on that conversation, 1213 01:02:08,360 --> 01:02:09,920 make predictions about who they are. 1214 01:02:10,080 --> 01:02:12,520 And those predictions are far more accurate than anything 1215 01:02:12,520 --> 01:02:15,040 we've had before, which makes a lot of sense, right? 1216 01:02:15,040 --> 01:02:17,040 Because if the chat bot figures out, well, I don't really know 1217 01:02:17,040 --> 01:02:19,640 enough about his level of contentiousness, can just ask 1218 01:02:19,640 --> 01:02:21,360 questions that go in this direction. 1219 01:02:21,640 --> 01:02:24,920 So it's like almost a self-guided dynamic scouting of 1220 01:02:24,920 --> 01:02:26,640 someone's personality. So that's Part 1. 1221 01:02:26,640 --> 01:02:30,520 Part 1 is that we can now democratize and scale the 1222 01:02:30,520 --> 01:02:33,440 prediction of psychological traits just by having these 1223 01:02:33,440 --> 01:02:36,520 tools available. But that's obviously only part 1224 01:02:36,520 --> 01:02:40,280 of the story because it also generates content, right? 1225 01:02:40,280 --> 01:02:44,040 So if the moment that I know is amazing at generating content, 1226 01:02:44,040 --> 01:02:47,040 that's why it's called AI, believe it or not. 1227 01:02:47,560 --> 01:02:51,280 But so once you give it like the instructions to say, well, 1228 01:02:51,520 --> 01:02:54,200 create me for this is like an example that we used in one of 1229 01:02:54,200 --> 01:02:57,120 our studies, come up with a short political speech 1230 01:02:57,440 --> 01:03:00,560 advocating for climate action for someone who scores high on 1231 01:03:00,560 --> 01:03:04,200 the moral foundation of fairness, which is like, again, 1232 01:03:04,200 --> 01:03:06,120 the way that people think about what's right or wrong in the 1233 01:03:06,120 --> 01:03:08,920 world, often times doesn't happen in the Big 5 personality 1234 01:03:08,920 --> 01:03:11,200 space, but more in the space of moral foundations. 1235 01:03:11,520 --> 01:03:14,640 And it does a remarkable job. So I don't need my team of 1236 01:03:14,640 --> 01:03:15,520 creed. That's right. 1237 01:03:15,520 --> 01:03:19,120 When I started, I had to head up a creative team that sits down, 1238 01:03:19,120 --> 01:03:22,520 comes up with a pictures, comes up with the text that goes on 1239 01:03:22,520 --> 01:03:24,640 the ad. And that was extremely limiting. 1240 01:03:24,640 --> 01:03:27,960 So with a beauty retailer example that I gave earlier, we 1241 01:03:27,960 --> 01:03:30,680 only tailored the ad that people saw on Facebook. 1242 01:03:30,680 --> 01:03:33,400 The moment that they went to the website, it looked the same for 1243 01:03:33,400 --> 01:03:35,240 everybody because we just clearly didn't have the 1244 01:03:35,240 --> 01:03:38,800 resources to now customize every single page on the beauty 1245 01:03:38,800 --> 01:03:41,800 retailer's website. Could be totally done today. 1246 01:03:41,880 --> 01:03:44,280 At the moment that I know that you're extroverted, we could 1247 01:03:44,920 --> 01:03:48,240 just in time create an entire website for you that's 1248 01:03:48,240 --> 01:03:51,800 specifically tailored both to who you are, maybe the context, 1249 01:03:51,800 --> 01:03:54,080 maybe the steps that you've already taken on the website. 1250 01:03:54,360 --> 01:03:57,440 And that's a total game changer. Yeah. 1251 01:03:57,440 --> 01:03:59,120 And, and when you're personalizing for moral 1252 01:03:59,120 --> 01:04:02,360 foundations, what's, what are some examples that ChatGPT would 1253 01:04:02,360 --> 01:04:06,160 come up with when tailoring a message to say for climate 1254 01:04:06,320 --> 01:04:09,640 action for a conservative versus a liberal? 1255 01:04:09,640 --> 01:04:11,960 Yeah. So I mean, So what we know is 1256 01:04:11,960 --> 01:04:15,840 that the moral foundations that are more closely and lined with 1257 01:04:15,840 --> 01:04:17,440 being liberal are fairness and care. 1258 01:04:17,680 --> 01:04:20,480 So when you pop that into ChatGPT, it just essentially 1259 01:04:20,480 --> 01:04:23,120 talks about how it's our responsibility to care for the 1260 01:04:23,120 --> 01:04:26,440 planet because it's not just us, and it's only fair that we take 1261 01:04:26,440 --> 01:04:29,280 care because future generations are otherwise going to suffer. 1262 01:04:29,560 --> 01:04:31,160 It's a very liberal argument, right? 1263 01:04:31,160 --> 01:04:34,320 It kind of we have a responsibility to look out for 1264 01:04:34,320 --> 01:04:36,440 others, even though we're not going to be impacted 1265 01:04:36,480 --> 01:04:39,520 immediately. Potentially the the arguments 1266 01:04:39,520 --> 01:04:42,960 for for more conservative people are usually focused on purity 1267 01:04:42,960 --> 01:04:46,920 and authority and loyalty. So if you think about loyalty is 1268 01:04:46,920 --> 01:04:50,400 like essentially we have this connection to Mother Earth and 1269 01:04:50,400 --> 01:04:53,000 there's something that like the loyalty that we show towards the 1270 01:04:53,000 --> 01:04:55,640 people in our community and mother is something that we 1271 01:04:55,640 --> 01:04:58,640 should be focused on or we just want to have to make sure that 1272 01:04:58,640 --> 01:05:02,280 the planet is in pristine shape. That's kind of playing. 1273 01:05:02,880 --> 01:05:04,200 Not so much for future generations. 1274 01:05:04,200 --> 01:05:05,520 Exactly. Not so much for future 1275 01:05:05,520 --> 01:05:07,040 generation. That's because we care about 1276 01:05:07,040 --> 01:05:08,560 other people. But that's just like 1. 1277 01:05:09,640 --> 01:05:13,560 We just have to be pure in the way that we act and in the way 1278 01:05:13,560 --> 01:05:14,840 that we interact with the planet. 1279 01:05:14,840 --> 01:05:18,200 So I think those are like, you just have to look at them. 1280 01:05:18,200 --> 01:05:21,360 And again, just take any personality or any psychological 1281 01:05:21,360 --> 01:05:24,080 trait that you can come up with and it just does a really 1282 01:05:24,080 --> 01:05:26,520 remarkable job. So I find it fascinating just 1283 01:05:26,520 --> 01:05:29,120 playing around with it and seeing what it comes up with. 1284 01:05:29,120 --> 01:05:31,480 And it's often times so much better than anything I would 1285 01:05:31,480 --> 01:05:33,680 have come up with. Far better. 1286 01:05:34,880 --> 01:05:40,120 And I guess I want to ask you in terms of what I felt recently in 1287 01:05:40,120 --> 01:05:44,680 the last few years, you talked about science fiction and the 1288 01:05:44,680 --> 01:05:47,240 difference and just being timing. 1289 01:05:47,240 --> 01:05:50,240 And a lot of these things would be science fictions before, but 1290 01:05:50,240 --> 01:05:53,320 now it's kind of more used to science or like reality. 1291 01:05:53,600 --> 01:05:57,680 And for me, I ended up used for fun creating, you know, AI 1292 01:05:57,680 --> 01:06:00,520 clones of my voice and my face and all of these things. 1293 01:06:00,960 --> 01:06:03,760 And through this kind of automation as well, you know, 1294 01:06:03,840 --> 01:06:08,080 the end point could be kind of a short video message from me that 1295 01:06:08,080 --> 01:06:09,840 looks like it's for me, sounds like for me. 1296 01:06:10,040 --> 01:06:13,000 My partner could. Not really see that it's not me, 1297 01:06:13,800 --> 01:06:16,920 but it's just something mass generated thing that it's just 1298 01:06:16,920 --> 01:06:21,960 starting from a simple, let's say and starting point of some 1299 01:06:21,960 --> 01:06:26,040 data point. It flows into kind of a 22nd 1300 01:06:26,320 --> 01:06:29,840 message talking about the person's not only name a certain 1301 01:06:29,840 --> 01:06:33,120 things, but also you like in a way that is likely to maybe make 1302 01:06:33,120 --> 01:06:35,960 them feel seen or, or appreciated or or heard and so 1303 01:06:36,000 --> 01:06:38,760 on. And I think that being possible 1304 01:06:38,840 --> 01:06:42,320 is science fiction. Like I think it's so it's 1305 01:06:42,360 --> 01:06:44,760 strange for me. I mean, I think it's, it's, it's 1306 01:06:44,760 --> 01:06:47,440 a terrifying, I think science, right, because I don't think 1307 01:06:47,440 --> 01:06:49,480 it's science fiction, right. So you can totally do that. 1308 01:06:49,480 --> 01:06:52,160 As you said, you, you didn't, the question is like, how good 1309 01:06:52,160 --> 01:06:55,280 is it? And for me, the, I was recently 1310 01:06:55,280 --> 01:06:58,480 talking to a New York Times journalist who outsourced her 1311 01:06:58,480 --> 01:07:00,960 decision making for a week. It's an amazing article if you 1312 01:07:00,960 --> 01:07:03,200 want to read it, But essentially she outsourced her decision 1313 01:07:03,200 --> 01:07:05,320 making for a week to just generative AI. 1314 01:07:05,840 --> 01:07:08,240 And, and one of the things that we talked about, which then she 1315 01:07:08,240 --> 01:07:12,680 also describes in the article is that it's really helpful in many 1316 01:07:12,680 --> 01:07:14,240 ways, right? Because it kind of has this in 1317 01:07:14,240 --> 01:07:16,560 like it can give you cooking instructions much better than 1318 01:07:16,560 --> 01:07:18,120 you could kind of get from the Internet. 1319 01:07:18,480 --> 01:07:21,920 But what it really does it she says, like it turns it turned me 1320 01:07:21,920 --> 01:07:25,400 into a basic bitch. And that was like one of the one 1321 01:07:25,400 --> 01:07:28,280 of the concerns that I have is that it's just going to make us 1322 01:07:28,280 --> 01:07:31,680 incredibly boring, right? So I think AI is really good as 1323 01:07:31,680 --> 01:07:35,840 a society, homogeneous, as a society and also as individuals, 1324 01:07:35,840 --> 01:07:38,120 because it's just kind of playing into the same thing. 1325 01:07:38,120 --> 01:07:40,720 And again, right. So if the things that Samuel is 1326 01:07:41,640 --> 01:07:46,480 an extrovert, that's how it's going to talk with everybody 1327 01:07:46,480 --> 01:07:49,240 that you're interacting with and then the feedback that you get 1328 01:07:49,240 --> 01:07:51,240 is just kind of reinforcing everything that you've done. 1329 01:07:51,440 --> 01:07:55,240 So I think that the way that we kind of create these 1330 01:07:55,240 --> 01:07:59,120 doppelgangers of AI, if we don't bake some exploration and some 1331 01:07:59,120 --> 01:08:02,240 uniqueness into the way that we designed them, we're just all 1332 01:08:02,240 --> 01:08:04,880 becoming like so average and so boring. 1333 01:08:04,880 --> 01:08:07,680 I think that's one of my when I'm concerned about privacy, the 1334 01:08:07,680 --> 01:08:11,040 loss of self determination, but I'm also really concerned about 1335 01:08:11,040 --> 01:08:14,640 just us being so boring. Have you there is this game, I 1336 01:08:14,640 --> 01:08:18,160 don't know if it's a game that that people have been sharing 1337 01:08:18,160 --> 01:08:22,000 where you actually, if you are a regular user of ChatGPT, you can 1338 01:08:22,000 --> 01:08:26,160 go to ChatGPT and ask it to tell you something about yourself 1339 01:08:26,399 --> 01:08:28,920 that you don't know. Have you experimented with I've? 1340 01:08:28,920 --> 01:08:30,960 Not done that yet. Oh. 1341 01:08:31,439 --> 01:08:34,560 Very funny because you, you need, I mean ChatGPT would need 1342 01:08:34,600 --> 01:08:37,920 a bunch of information about me to play the game well, right? 1343 01:08:37,920 --> 01:08:41,439 So it's an interesting one. Talking about games. 1344 01:08:43,040 --> 01:08:46,279 Quick, quick game we have for this season of the podcast is 1345 01:08:46,279 --> 01:08:48,560 something called To AI or Not to AI? 1346 01:08:48,760 --> 01:08:51,720 And this will be a very quick fire round where basically we'll 1347 01:08:51,720 --> 01:08:54,920 give you a few prompts and you used to have to answer what do 1348 01:08:54,920 --> 01:08:58,080 you think it's a good suited or not well suited for AI? 1349 01:08:58,880 --> 01:09:02,240 To outsource to AI, Yeah, well, we'll see what the questions are 1350 01:09:02,960 --> 01:09:03,560 exactly. Good. 1351 01:09:04,160 --> 01:09:05,880 OK. Are you ready? 1352 01:09:06,200 --> 01:09:08,680 Yep. Clean your room. 1353 01:09:09,439 --> 01:09:11,359 Oh, if there was like a physical AI? 1354 01:09:11,359 --> 01:09:16,080 Absolutely. What about make sure you wake up 1355 01:09:16,080 --> 01:09:21,880 on time? Yeah, AI, you're like, yes, 1356 01:09:21,880 --> 01:09:26,800 everything generate targeted ads for potentially pregnant 1357 01:09:26,800 --> 01:09:30,600 customers. I mean, is it counterfactual to 1358 01:09:30,600 --> 01:09:34,080 have humans generate targeted ads for potentially pregnant 1359 01:09:34,080 --> 01:09:35,080 customers? No the counter. 1360 01:09:35,080 --> 01:09:37,240 No the counter. This is a task that is well 1361 01:09:37,240 --> 01:09:40,680 suited for AI and doesn't have to be an alternative. 1362 01:09:41,760 --> 01:09:44,120 Well, I mean, I think AI would do a remarkable job. 1363 01:09:44,120 --> 01:09:45,960 Would I want to be targeted by it? 1364 01:09:45,960 --> 01:09:48,960 That depends on whether I ask for it or not. 1365 01:09:49,040 --> 01:09:50,760 Because do I think it could actually give me amazing 1366 01:09:50,760 --> 01:09:51,920 recommendations? Yes. 1367 01:09:52,359 --> 01:09:53,920 Do I want to do this behind my back? 1368 01:09:53,920 --> 01:09:55,480 Probably not. With consent, yes. 1369 01:09:57,480 --> 01:10:00,960 What about improve the Big 5 model of personality? 1370 01:10:01,320 --> 01:10:06,040 Yeah, great. Predict your ideal pet. 1371 01:10:07,880 --> 01:10:10,280 Or if I don't have to follow that recommendation, I'll take 1372 01:10:10,280 --> 01:10:15,440 the idea. OK, OK, so final one, 1373 01:10:16,160 --> 01:10:19,720 personalize the way your GPS speaks to you based on your 1374 01:10:19,720 --> 01:10:24,800 personality profile. Personalize the way that my GPS 1375 01:10:24,800 --> 01:10:29,320 so my navigation in the car is that what you and yeah, I don't 1376 01:10:29,320 --> 01:10:32,400 have a have an issue with that. I think what is true for all of 1377 01:10:32,400 --> 01:10:35,480 these questions is that I would want to understand how it works 1378 01:10:35,480 --> 01:10:37,520 and what it does. And most of the time, I think 1379 01:10:37,520 --> 01:10:40,520 once you understand it and you have some control, it can be 1380 01:10:40,520 --> 01:10:43,120 extremely helpful. So that's, I think generally 1381 01:10:43,120 --> 01:10:47,080 speaking, I err on the side of there's a huge opportunity if we 1382 01:10:47,080 --> 01:10:51,160 manage it well and less so on the we're all doomed. 1383 01:10:53,480 --> 01:10:56,960 Well, that's a perfect segue to our final question that we ask 1384 01:10:57,080 --> 01:10:59,280 all of our guests. What is your most controversial 1385 01:10:59,280 --> 01:11:03,000 opinion about AI? That's I mean that that might be 1386 01:11:03,000 --> 01:11:04,640 it. It's just funny because I don't 1387 01:11:04,680 --> 01:11:07,600 think it's really hard to have controversial opinions of AI 1388 01:11:07,600 --> 01:11:10,200 because like people's opinions are all over the place, right. 1389 01:11:10,200 --> 01:11:11,400 So you. Can get all the opinions are 1390 01:11:11,400 --> 01:11:12,840 already out. There, Yeah, they're already out 1391 01:11:12,840 --> 01:11:14,800 there. But I do think that I probably, 1392 01:11:14,800 --> 01:11:18,680 if you kind of compare me to the average person in that space, I 1393 01:11:18,680 --> 01:11:20,080 think I'm probably more optimistic. 1394 01:11:20,560 --> 01:11:23,600 I think that where I'm less pessimistic is that we have the 1395 01:11:23,600 --> 01:11:25,560 ability to do it ourselves as users. 1396 01:11:25,560 --> 01:11:28,360 So I don't think we can just have regulations and mandate 1397 01:11:28,360 --> 01:11:31,160 transparency and control and say, well, would just tell you 1398 01:11:31,160 --> 01:11:33,600 what to do and then you can navigate the landscape yourself. 1399 01:11:33,600 --> 01:11:36,640 That I'm very pessimistic about. But I do think that there's so 1400 01:11:36,640 --> 01:11:39,640 many good use cases and with just like the right regulation 1401 01:11:39,640 --> 01:11:42,200 and some of the new technologies that allow you to do it in a 1402 01:11:42,200 --> 01:11:45,920 privacy preserving way and give you more control in a supported 1403 01:11:45,920 --> 01:11:48,920 way, then I think the opportunities are incredible. 1404 01:11:50,480 --> 01:11:53,320 Yeah, you, you talk in your book about having better systems by 1405 01:11:53,320 --> 01:11:54,120 design. Yeah. 1406 01:11:54,120 --> 01:11:58,120 Making, not putting that burden on the user to, for example, 1407 01:11:58,120 --> 01:12:00,480 read the terms and conditions and know what they're signing up 1408 01:12:00,480 --> 01:12:04,000 for, but rather help them without all of that effort. 1409 01:12:04,600 --> 01:12:05,440 Exactly. So awesome. 1410 01:12:05,720 --> 01:12:07,880 Yeah, this is so much fun. I feel like we could have talked 1411 01:12:07,880 --> 01:12:11,200 for so much more. But I guess like for anyone who 1412 01:12:11,200 --> 01:12:12,960 feels like the one to learn more. 1413 01:12:13,640 --> 01:12:17,840 There's a book they can read, so we will make sure to help people 1414 01:12:17,840 --> 01:12:20,040 along the way to to get the book and. 1415 01:12:20,120 --> 01:12:22,640 Yeah, it's just out. Yeah, it's just out. 1416 01:12:22,640 --> 01:12:25,680 And it's funny because I, I didn't start writing it with a 1417 01:12:25,960 --> 01:12:27,720 book in mind. So I just started writing these 1418 01:12:28,000 --> 01:12:31,320 short stories from my friends and family just to kind of give 1419 01:12:31,320 --> 01:12:33,520 the well, first of all, to make it more enjoyable to follow my 1420 01:12:33,520 --> 01:12:36,320 research because I don't think reading scientific articles is 1421 01:12:36,320 --> 01:12:39,040 just what they need, but also just to tell them like, look, 1422 01:12:39,040 --> 01:12:40,360 here's what your data says about you. 1423 01:12:40,360 --> 01:12:42,400 Here's why maybe you should care about it as well. 1424 01:12:42,400 --> 01:12:44,920 So I try to make it as a approachable and accessible as 1425 01:12:44,920 --> 01:12:47,720 as possible. It's amazing. 1426 01:12:48,000 --> 01:12:50,200 And yeah, it was lovely to read it. 1427 01:12:50,200 --> 01:12:54,160 And yeah, so congratulations on on the book and thanks so much 1428 01:12:54,160 --> 01:12:56,880 for coming on today. It was, yeah, fantastic. 1429 01:12:56,880 --> 01:12:59,600 Two foundations to speak, and. Yeah, thanks for all your 1430 01:12:59,600 --> 01:13:01,800 wonderful work and your research and and everything. 1431 01:13:01,800 --> 01:13:04,200 It's fantastic. Thanks for having me, was really 1432 01:13:04,200 --> 01:13:08,280 fun. And that's a wrap. 1433 01:13:08,720 --> 01:13:11,320 You've been listening to the Behavioral Design Podcast 1434 01:13:11,600 --> 01:13:14,120 brought to you by Habit Weekly and Nuanced Behavior. 1435 01:13:14,520 --> 01:13:16,800 Sam and Alene tell me. This season is packed with 1436 01:13:16,800 --> 01:13:20,680 incredible insights about behavioral design and AI, so be 1437 01:13:20,680 --> 01:13:23,240 sure to subscribe and share the podcast with your friends. 1438 01:13:23,480 --> 01:13:25,680 Though you might want to keep it away from your enemies. 1439 01:13:27,000 --> 01:13:29,640 In case you haven't noticed, I'm an AI voice. 1440 01:13:30,800 --> 01:13:33,600 Yep, pretty crazy. Quite the improvement since last 1441 01:13:33,600 --> 01:13:35,600 season's AI outro, don't you think? 1442 01:13:36,840 --> 01:13:39,520 If you'd like to collaborate with us at Nuance Behavior, 1443 01:13:39,720 --> 01:13:42,440 where we use behavioral design to craft digital products with 1444 01:13:42,440 --> 01:13:46,760 Nuance, e-mail us at hello@nuancebehavior.com or book 1445 01:13:46,760 --> 01:13:50,000 a call directly on our website, nuancebehavior.com. 1446 01:13:51,440 --> 01:13:54,880 A special thanks to the amazing Dave Pizarro for our show music 1447 01:13:55,080 --> 01:13:58,000 and to Mei Chen Yap and April English for their help in 1448 01:13:58,000 --> 01:13:59,920 producing and publishing this episode. 1449 01:14:00,360 --> 01:14:03,200 Thanks again for tuning in. We'll be back soon with another 1450 01:14:03,200 --> 01:14:06,400 exciting conversation where behavioral design and AI 1451 01:14:06,400 --> 01:14:14,320 intersect. Happens to Murgatroyd. 1452 01:14:21,480 --> 01:15:08,200 Oh. Well, hello there. 1453 01:15:08,200 --> 01:15:10,720 Are you still listening? Wow, we really. 1454 01:15:10,800 --> 01:15:12,320 Why? Yeah, why? 1455 01:15:12,320 --> 01:15:14,440 Why are you still here? Why are you still listening? 1456 01:15:14,440 --> 01:15:16,200 Get out of here. Get out of here. 1457 01:15:16,200 --> 01:15:20,760 But I guess to reward you for being such a loyal listener and 1458 01:15:20,840 --> 01:15:24,640 making it all the way to our kind of like little hidden gem 1459 01:15:24,640 --> 01:15:27,640 of a final snippet, we have a recommendation. 1460 01:15:28,640 --> 01:15:32,360 If you liked this episode with Sandra, check out our episode on 1461 01:15:32,360 --> 01:15:35,000 personality with Sanjay Servastava. 1462 01:15:35,080 --> 01:15:38,120 He is one of my former colleagues at Apple. 1463 01:15:38,160 --> 01:15:42,520 In that episode, we got all the way into the details of 1464 01:15:42,520 --> 01:15:46,120 personality, how it's measured, how it's used and so on. 1465 01:15:46,120 --> 01:15:47,160 So. Check it out. 1466 01:15:48,200 --> 01:15:54,040 Yeah, so sorry, no embarrassing L takes this one but. 1467 01:15:57,320 --> 01:16:00,080 I know, I thought it was so ironic that you said that as 1468 01:16:00,080 --> 01:16:05,480 we're like stumbling over. OK, we'll just say. 1469 01:16:07,840 --> 01:16:09,720 I don't think you have to say anything else. 1470 01:16:10,240 --> 01:16:11,600 OK. I think we can end with that. 1471 01:16:11,600 --> 01:16:13,480 We'll. End with that only if you want 1472 01:16:13,480 --> 01:16:16,000 to bye bye bye.