WEBVTT - Sleepwalking 0:00:01.760 --> 0:00:04.200 Sleep Walkers is a production of I Heart Radio and 0:00:04.360 --> 0:00:10.000 Unusual Productions. So I mean, let's just ground the conversation 0:00:10.039 --> 0:00:13.239 for a second. There's two point two billion people who 0:00:13.360 --> 0:00:16.640 use Facebook. It's about the size of Christianity. There's one 0:00:16.640 --> 0:00:19.040 point nine billion people who use YouTube. It's about the 0:00:19.079 --> 0:00:22.520 size of Islam. That's Tristan Harris. When I walk into 0:00:22.560 --> 0:00:25.520 that Facebook room, you know, I'm walking into a room 0:00:25.560 --> 0:00:28.760 that's designed to never make me leave with a thousand 0:00:28.840 --> 0:00:32.159 engineers using supercomputers to calculate the perfect seductive thing to 0:00:32.200 --> 0:00:34.920 put in front of my brain. And Tristan knows a 0:00:35.000 --> 0:00:38.120 thing or two about seduction. I was a design ethosistic 0:00:38.200 --> 0:00:41.199 Google and have spent a decade understanding some of the 0:00:41.240 --> 0:00:44.320 invisible forces that shape the way that we see and 0:00:44.360 --> 0:00:46.360 make sense of the world and choose in the world. 0:00:46.640 --> 0:00:50.240 In case you're wondering they teach this stuff in school, well, 0:00:50.440 --> 0:00:53.200 they do at Stanford. I studied at this lab called 0:00:53.200 --> 0:00:57.480 the Stanford Persuasive Technology Lab that teaches engineering students about 0:00:57.480 --> 0:01:00.680 the entire discipline of persuasion. Everything you know how to 0:01:00.720 --> 0:01:03.040 persuade dogs with clicker training, So click click, you know 0:01:03.080 --> 0:01:05.280 you've got the food. Click click, reward all the way 0:01:05.280 --> 0:01:07.440 to casino design and slot machines, and how do you 0:01:07.560 --> 0:01:09.520 change the lighting to get people to buy things? And 0:01:09.560 --> 0:01:12.319 then you know supermarket design and choice architecture and putting 0:01:12.360 --> 0:01:14.320 the candy in the final aisle, because that's the thing 0:01:14.360 --> 0:01:17.920 that gets you to buy. Tristan and his classmates used 0:01:17.959 --> 0:01:21.280 what they learned at the Persuasive Technology Lab to define 0:01:21.319 --> 0:01:24.720 how we live online. My partners in that class, and 0:01:24.800 --> 0:01:27.560 two thousands of sex at Stanford were the founders of Instagram. 0:01:27.680 --> 0:01:29.679 Half my friends built some of these products and then 0:01:29.720 --> 0:01:32.319 to the like button. Think about that for a moment. 0:01:32.840 --> 0:01:35.360 The people who designed the apps we use the most 0:01:35.840 --> 0:01:38.880 learned how to make them as appealing as possible by 0:01:38.920 --> 0:01:42.000 borrowing the science that trains dogs and gets people hooked 0:01:42.040 --> 0:01:44.800 on gambling. And now all of that is baked into 0:01:44.800 --> 0:01:48.120 a device that's basically become an extension of our body. 0:01:48.480 --> 0:01:51.040 We check our phones about eighty times a day, that's 0:01:51.040 --> 0:01:54.320 the conservative number. And you know their incentive is to 0:01:54.480 --> 0:01:57.960 calculate what is the perfect most seductive thing I can 0:01:57.960 --> 0:02:00.040 show you next to the most seductive red color for 0:02:00.120 --> 0:02:03.280 that notification, or the most seductive video that you know 0:02:03.440 --> 0:02:06.520 you can't help but want to watch next. At a 0:02:06.600 --> 0:02:09.480 time when technology is changing faster than our ability to 0:02:09.560 --> 0:02:13.200 understand it and seeping into nearly every corner of our lives, 0:02:13.560 --> 0:02:17.000 what kind of murky future are we sleepwalking into? What 0:02:17.040 --> 0:02:19.240 can we do to take back some control? And how 0:02:19.360 --> 0:02:23.400 is emerging technology changing our lives for the better. This 0:02:24.040 --> 0:02:39.720 is Sleepwalkers, So welcome. I'mrs and I'm doing this show 0:02:39.720 --> 0:02:42.440 because I'm fascinated with how we relate to the technologies 0:02:42.480 --> 0:02:45.000 that they're changing our lives, whether they end up being 0:02:45.040 --> 0:02:49.280 something like Dr Frankenstein's Monster or ai craft it seltzer Water. 0:02:49.840 --> 0:02:52.200 In this episode, we look at how our technology gets 0:02:52.200 --> 0:02:56.320 into our heads and we take on some distinctly modern monsters, 0:02:56.680 --> 0:03:01.080 from the trick to successful online dating to deterring potential 0:03:01.240 --> 0:03:05.919 terrorists with invisible technology. And I'll have some company along 0:03:05.919 --> 0:03:10.680 the way. Hi. That's Kara Price. She's my friend and 0:03:10.720 --> 0:03:13.000 she hosted a show called Talk Nerdy to Me for 0:03:13.040 --> 0:03:15.840 the Huffington Post. Um. So, I saw this article on 0:03:15.880 --> 0:03:19.160 the Times, the New York Times about how all of 0:03:19.200 --> 0:03:23.320 these Silicon Valley exacts are taking away their kids screen time, 0:03:23.840 --> 0:03:26.679 you know, telling their kids nanny's that the nanny can't 0:03:26.800 --> 0:03:30.799 use screens. And there's actually this quote from Mark Zuckerberg's 0:03:30.800 --> 0:03:35.640 former executive assistant, this woman Athena Tivaria. She says, I 0:03:35.680 --> 0:03:38.200 am convinced the devil lives in our phones and is 0:03:38.240 --> 0:03:41.520 wreaking havoc on our children. Is that like the devil 0:03:41.600 --> 0:03:43.360 is in the details, or is it like the devil 0:03:43.360 --> 0:03:45.320 who lives in hell? No? I think it's the red 0:03:45.560 --> 0:03:48.240 devil who wears no clothes that guys in the phone. Yeah, 0:03:48.240 --> 0:03:50.920 because you know what he does. It leaves people in temptation. 0:03:51.200 --> 0:03:53.280 That's right, that's the devil's show. That's really right, and 0:03:53.320 --> 0:03:55.040 it is have you ever seen a child on an iPad? 0:03:55.120 --> 0:03:58.600 Have you ever seen me? Well, have you ever seen me? Yes? Mesmerized, 0:03:58.600 --> 0:04:01.600 clickity clack, don't come back, actually showing Pocca, who's the 0:04:01.640 --> 0:04:05.640 first president of Facebook, said God only knows what it's 0:04:05.680 --> 0:04:09.400 doing to our children's brains. Got on the devil. Here's 0:04:09.440 --> 0:04:14.040 all of these technology executives who have built basically what 0:04:14.080 --> 0:04:17.440 we use every day to do everything from being in 0:04:17.440 --> 0:04:20.240 touch with our friends to meeting our loves to getting 0:04:20.240 --> 0:04:23.320 from A to B. And suddenly they're saying, well, you guys, 0:04:23.320 --> 0:04:26.919 go ahead, but not for my kids. Yeah, yeah. I 0:04:26.920 --> 0:04:29.880 think it's interesting that this feels very similar to the 0:04:29.880 --> 0:04:33.839 conversation we have surrounding you know, junk food, sugar, tobacco, alcohol, 0:04:33.880 --> 0:04:36.160 that these are things that your parents are supposed to 0:04:36.320 --> 0:04:39.280 protect you from up until a certain point, whether it's 0:04:39.320 --> 0:04:42.039 when you go to college or whatever. And now our 0:04:42.080 --> 0:04:44.200 technology used, whether it be how much we're on social 0:04:44.200 --> 0:04:47.760 media or how much time we're spending gaming on our screens, 0:04:48.240 --> 0:04:51.000 is something that parents have to regulate. The problem is 0:04:51.040 --> 0:04:53.400 one like I'm too old for my parents to regulate it, 0:04:53.440 --> 0:04:55.520 so so what am I going to do? And also 0:04:55.560 --> 0:04:57.520 like my parents use it, so it is sort of 0:04:57.520 --> 0:04:59.279 like sugar in the sense it's like there are parents 0:04:59.279 --> 0:05:01.000 that tell their kids not each sugar who have major 0:05:01.040 --> 0:05:03.919 sugar addictions. We had those conversations in the past, and 0:05:03.920 --> 0:05:06.280 now we're starting to have them about technology. But how 0:05:06.279 --> 0:05:09.080 do we tell our kids to regulate their use of technology? 0:05:09.080 --> 0:05:11.960 And we can't regulate our own And wants to say 0:05:11.960 --> 0:05:17.400 to us that we can't regulate our own technology. His 0:05:17.560 --> 0:05:19.680 Cara with a story of what happens when you add 0:05:19.680 --> 0:05:23.640 service think they know everything about you, but actually get 0:05:23.680 --> 0:05:30.120 it all wrong. I just wrote the whole thing in 0:05:30.200 --> 0:05:36.640 like thirty minutes just you know, banging the keyboard, just 0:05:37.040 --> 0:05:41.240 sobbing the whole time. Dear tech companies. Gillian is a 0:05:41.320 --> 0:05:44.400 social media power user. She uses Twitter for work at 0:05:44.400 --> 0:05:47.919 the Washington Post and Facebook for her social calendar. Even 0:05:47.920 --> 0:05:51.240 her wedding is on YouTube, and in late she was 0:05:51.360 --> 0:05:53.800 entering the final months of her pregnancy. She and her 0:05:53.880 --> 0:05:56.560 husband Bobby were doing the normal what to expect when 0:05:56.560 --> 0:06:00.000 you're expecting stuff, preparing for their newborn son to come 0:06:00.000 --> 0:06:04.680 into their lives. I can remember about two days before 0:06:04.920 --> 0:06:10.000 everything happened, soaking in the tub, and I was just thinking, God, 0:06:10.120 --> 0:06:14.880 this has just gone perfectly, Like I've never been accidentally 0:06:14.960 --> 0:06:20.359 pregnant despite not always being responsible, and we had no 0:06:20.520 --> 0:06:24.320 problem getting pregnant when we wanted to, despite my being 0:06:24.400 --> 0:06:27.919 thirty eight. So I was just like, wow, one and done. 0:06:28.400 --> 0:06:32.600 This totally worked. A couple of days later, I was 0:06:32.760 --> 0:06:36.479 feeling some pain, so I called the doctor's office and 0:06:36.520 --> 0:06:40.240 I said, well, I haven't felt a move today, and 0:06:40.360 --> 0:06:42.719 they said, okay, come in, and so I went in 0:06:42.839 --> 0:06:46.920 and they put me on the sonogram, and I knew immediately. 0:06:48.200 --> 0:06:55.360 I was just screaming no, no, no. A week or 0:06:55.400 --> 0:06:58.240 so after Gillian's sun died. She was being haunted by 0:06:58.279 --> 0:07:01.520 targeted advertising ad that assumed she had given birth to 0:07:01.560 --> 0:07:05.640 a healthy baby boy. When I would scroll through Facebook 0:07:05.760 --> 0:07:09.920 or Instagram, I would get maternity where ads, and so 0:07:09.960 --> 0:07:11.440 I was like, okay, I have to teach it that 0:07:11.480 --> 0:07:14.880 I'm not pregnant anymore. Gillian did the only thing people 0:07:14.960 --> 0:07:17.560 can do with ads they don't want to see. She 0:07:17.720 --> 0:07:19.640 clicked the three dots in the corner of the ad 0:07:19.760 --> 0:07:22.760 and gave feedback. So I would say I don't want 0:07:22.760 --> 0:07:25.960 to see this ad, and then it would say why, 0:07:26.080 --> 0:07:28.880 and I would say, because it's not relevant to me, 0:07:29.080 --> 0:07:32.920 which is like so hard to acknowledge. But Gillian learned 0:07:32.920 --> 0:07:35.520 the hard way that the Facebook algorithm isn't programmed for 0:07:35.560 --> 0:07:38.160 the outcome of a still birth. Then when I god, 0:07:38.200 --> 0:07:41.160 when I got that experienced email, I just I can't 0:07:41.200 --> 0:07:45.600 even like I just snapped finished registering your baby for 0:07:45.920 --> 0:07:49.440 lifelong credit tracking. I just I was just like, you 0:07:49.520 --> 0:07:53.840 have got to be kidding me. Gillian was angry, and 0:07:53.960 --> 0:07:56.200 she knew she had a platform, so she wrote the 0:07:56.280 --> 0:07:59.600 letter and posted it on Twitter. Dear tech companies, I 0:07:59.640 --> 0:08:02.760 know you knew I was pregnant. You probably saw me 0:08:02.840 --> 0:08:07.000 googling maternity plaid and baby safe crib paint, and I 0:08:07.000 --> 0:08:10.120 bet Amazon dot Com even told you my due date 0:08:10.320 --> 0:08:14.240 when I created that prime registry, and silly me, I 0:08:14.360 --> 0:08:16.960 even clicked once or twice on the maternity where ads 0:08:17.000 --> 0:08:20.440 Facebook served up. What can I say, I'm your ideal 0:08:20.560 --> 0:08:25.160 engaged user, but didn't you also see me googling baby 0:08:25.200 --> 0:08:28.720 not moving? And then the announcement posts with key where 0:08:28.720 --> 0:08:32.679 it's like heartbroken and problem and stillborn and the two 0:08:32.760 --> 0:08:37.439 hundred tear drop emoticons from my friends. Gillian's letters struck 0:08:37.440 --> 0:08:41.320 a chord. It got retweeted twenty eight thousand times, and 0:08:41.400 --> 0:08:44.600 shortly after Gillian posted her letter to Twitter, she received 0:08:44.600 --> 0:08:48.480 a response tweet from Facebook's VP of Ads, Rob Goldman. 0:08:49.440 --> 0:08:52.080 I'm so sorry for your loss, he said, and you're 0:08:52.120 --> 0:08:55.719 painful experience with our products. We have a setting available 0:08:55.720 --> 0:08:58.840 that can block ads about some topics people may find painful, 0:08:59.040 --> 0:09:02.760 including parenting. It still needs improvement, but please know that 0:09:02.800 --> 0:09:05.880 we're working on it and welcome your feedback. So I 0:09:05.920 --> 0:09:09.880 turned it off, and within a few hours I got 0:09:09.880 --> 0:09:15.240 an ad for adoption agencies, and the next day I 0:09:15.360 --> 0:09:20.840 got an ad for um father's son matching one sees. 0:09:21.360 --> 0:09:24.199 I've taken screenshots to while every time I get one, 0:09:24.240 --> 0:09:27.120 I just take a picture of it, just like it's 0:09:27.120 --> 0:09:31.440 not working. Yeah, the adoption one g f y. You know, 0:09:33.000 --> 0:09:35.520 Rob Goldman's advice may have changed the ads, but it 0:09:35.559 --> 0:09:39.880 didn't solve the fundamental problem. The algorithms couldn't stop reminding 0:09:39.880 --> 0:09:43.440 Gillian of her loss. Other people were like, well, don't 0:09:43.440 --> 0:09:45.800 be on Facebook at all, you know, don't do any 0:09:45.880 --> 0:09:48.720 of those things, And first of all, I don't think 0:09:48.720 --> 0:09:53.000 that's realistic, especially because for Gillian, Facebook was also providing 0:09:53.040 --> 0:09:55.920 comfort through it all. It was helpful to have my 0:09:56.040 --> 0:09:59.880 friends timing in and saying, we're so sorry, you know, 0:10:00.040 --> 0:10:02.400 what can we do? And you know, a bunch of 0:10:02.440 --> 0:10:04.920 people made an Uber eats fund for me and my 0:10:05.000 --> 0:10:07.400 husband so we could just have food delivered. While when 0:10:07.440 --> 0:10:09.959 we got home it's like, you know, the twenty one 0:10:10.040 --> 0:10:15.960 century bringing over lasagna. And this was all organized on Facebook. Yeah, 0:10:16.000 --> 0:10:19.040 it was on Facebook, and ironically it helped her connect 0:10:19.080 --> 0:10:21.480 with people going through the same thing. A woman who 0:10:21.559 --> 0:10:23.160 used to go to my church, she had a still 0:10:23.240 --> 0:10:26.120 birth on Christmas Day a few weeks after I did, 0:10:26.240 --> 0:10:30.599 and the memorial service for her baby was posted to Facebook, 0:10:30.960 --> 0:10:33.120 I wouldn't have known about it, and I wouldn't have 0:10:33.160 --> 0:10:39.360 gone if I hadn't checked Facebook. Gillian isn't planning to 0:10:39.360 --> 0:10:42.120 delete social media. She just wishes it could be better. 0:10:42.720 --> 0:10:44.440 She's still going to use it to keep in touch 0:10:44.480 --> 0:10:47.480 with friends and family. But whereas she was once comfortable 0:10:47.520 --> 0:10:50.400 with her wedding on YouTube, there are now something she 0:10:50.440 --> 0:10:53.440 won't be comfortable sharing. And I have to say, now 0:10:54.320 --> 0:10:57.880 having this experience, you know, I knew it was being tracked, 0:10:58.040 --> 0:11:01.199 but having the tracking revealed to me in such a 0:11:01.320 --> 0:11:05.559 garish display. If we do have a living child someday, 0:11:05.920 --> 0:11:08.079 I think that's going to be actually really easy for 0:11:08.120 --> 0:11:11.960 me to just be like, no Internet, you don't get 0:11:11.960 --> 0:11:24.280 to have that. Even after the experience. Gillian doesn't hate Facebook. 0:11:24.679 --> 0:11:27.960 At the same time, she wouldn't want her future children tracked, 0:11:28.360 --> 0:11:30.880 a bit like the Silicon Valley execs we talked about 0:11:30.880 --> 0:11:34.120 earlier taking their own children off social media. But what 0:11:34.240 --> 0:11:37.480 if the same targeting technology that harm Gillian could help 0:11:37.520 --> 0:11:41.040 others and keep people who are hurting from damaging themselves 0:11:41.040 --> 0:11:44.280 and society. And what if it wasn't a broad group 0:11:44.400 --> 0:11:47.600 like new mothers being targeted by ad bots, but a 0:11:47.640 --> 0:11:51.600 specific group of people having their personal search results changed. 0:11:52.160 --> 0:11:55.600 I heard about a programmer Alphabet, Google's parent company trying 0:11:55.640 --> 0:11:57.760 to do just that, and I wanted to know more. 0:11:58.600 --> 0:12:09.000 We'll get there after the brain. My name is Yasmin Green. 0:12:09.360 --> 0:12:13.119 I am the director of Research and Development at Jigsaw. 0:12:13.520 --> 0:12:16.360 We are sitting in one of our rooms in our 0:12:16.400 --> 0:12:19.360 New York office. The name of the meeting room is 0:12:19.600 --> 0:12:25.040 Smoke Signal Communication Technology. Yeah, they're all named after different 0:12:25.080 --> 0:12:29.280 communication technologies. Jigsaw is a part of Alphabet and Google 0:12:29.440 --> 0:12:34.719 that's focused on technology that addresses big security challenges. So 0:12:34.760 --> 0:12:37.559 how did Jigsaw begin? Two decades ago? When we were 0:12:37.559 --> 0:12:41.400 designing the platforms and apps, we were not really imagining 0:12:41.440 --> 0:12:46.600 that repressive governments and criminals and terrorist groups or like 0:12:46.679 --> 0:12:50.400 salivating and about innovating to use these platforms just as 0:12:50.440 --> 0:12:52.440 just as well as everyone else. Um. And now we 0:12:52.520 --> 0:12:55.120 realize that we really need to understand their goals and 0:12:55.160 --> 0:12:57.520 their activities if we want to keep people safe online. 0:12:57.880 --> 0:13:01.160 It's Yasmin's job to look at how those repressive regimes, 0:13:01.200 --> 0:13:05.200 criminals and terrorist groups operate, find trends and try to 0:13:05.320 --> 0:13:08.520 counter them. Our approaches to try to understand the human 0:13:08.600 --> 0:13:10.880 journeys and to understand the role of technology and see 0:13:10.920 --> 0:13:14.040 if we can build technology that stops people getting to 0:13:14.120 --> 0:13:16.640 the point that we would consider them of ident extremist 0:13:16.720 --> 0:13:20.040 or a terrorist. So if you want to understand terrorists, 0:13:20.320 --> 0:13:24.679 where do you start. Well, to better understand extremism, Yasmin's 0:13:24.720 --> 0:13:28.480 team use some somewhat extreme methods themselves. One of the 0:13:28.480 --> 0:13:30.720 first things we did in two thousand and eleven was 0:13:30.760 --> 0:13:35.959 we brought together eighty four former extremists and survivors of terrorism. 0:13:36.040 --> 0:13:38.199 And we had isn't a mist people who's going to 0:13:38.200 --> 0:13:43.320 find Afghanistan or al Shabab from Somalia, Nigerian isna miss militia, 0:13:43.400 --> 0:13:45.520 but we also had a form of violent Israeli settler. 0:13:45.640 --> 0:13:47.640 We had form of Christian militia. We've got them one 0:13:47.679 --> 0:13:51.760 together on one place. Can you imagine the security concern 0:13:52.000 --> 0:13:55.240 for Google in helping us convene everyone in one place. 0:13:55.280 --> 0:13:57.640 We had snipers on the roof, We had six months 0:13:57.679 --> 0:14:00.960 vetting for everyone. We cared about them being public and 0:14:00.960 --> 0:14:03.480 associated with an engine to bring people out of those 0:14:03.480 --> 0:14:07.480 extremist groups. And beyond gathering all the former terrorists in 0:14:07.559 --> 0:14:12.160 one place, Yasmin's team also invited the victims of terrorism. 0:14:12.240 --> 0:14:14.920 We had survivors, So we had people who celebrated nine 0:14:14.920 --> 0:14:17.640 eleven and people who had lost family members in eleven. 0:14:17.720 --> 0:14:21.560 We had a lady who said that she had woken 0:14:21.640 --> 0:14:23.800 up in hospital with a name band on her risk 0:14:23.880 --> 0:14:27.080 that said gender unknown because her body was like so mutilated. 0:14:27.600 --> 0:14:29.600 The point of bringing everyone together was to say what 0:14:29.920 --> 0:14:33.800 is common in the human radicalization path even when you're 0:14:33.800 --> 0:14:39.240 looking across ideologies, and does technology have a role to play? 0:14:39.440 --> 0:14:42.720 In other words, Yasmin wanted to understand how people become 0:14:42.920 --> 0:14:46.880 terrorists and what role the internet plays. But Chicksaw's role 0:14:46.960 --> 0:14:49.320 was never just to have summits and collect the information. 0:14:49.720 --> 0:14:53.000 They wanted to take action, and ISIS was the obvious 0:14:53.000 --> 0:14:56.480 place to start. In March of this year, ISIS actually 0:14:56.520 --> 0:14:59.680 lost the last of its remaining territory, but when the 0:15:00.040 --> 0:15:03.000 left fate was that it's height. Thousands of foreign fighters 0:15:03.080 --> 0:15:07.760 were recruited online. They left their homes in places like Germany, England, 0:15:07.840 --> 0:15:11.600 America for the battle fields of Syria. This trip that 0:15:11.640 --> 0:15:15.160 we took to Iraq, where we spoke to defectors of ISIS, 0:15:15.440 --> 0:15:17.560 was really instructive for us. We had these face to 0:15:17.600 --> 0:15:20.640 face conversations with these young men who had left home, 0:15:20.840 --> 0:15:24.360 gone to Syria, and Iraq trained with ISIS, did their 0:15:24.360 --> 0:15:27.560 religious indoctrination with ISIS, got their postings, some of them 0:15:27.560 --> 0:15:30.360 as suicide bombers, some of them as technical drivers, some 0:15:30.400 --> 0:15:33.560 of them as night watchmen. And they had realized that 0:15:33.640 --> 0:15:35.880 it was all a lie. So they're telling us about it, 0:15:35.920 --> 0:15:37.920 and we say, if you knew the day that you 0:15:38.000 --> 0:15:40.000 left everything that you know and now, would you still 0:15:40.080 --> 0:15:44.200 have gone? And they invariably said that they would still 0:15:44.240 --> 0:15:49.800 have gone. Honestly, I'd already finished, he asked me in 0:15:49.800 --> 0:15:51.960 a sentence. In my head, I thought she was going 0:15:52.000 --> 0:15:55.240 to say, they invariably said that they wouldn't have gone, 0:15:55.920 --> 0:15:59.840 but in fact, knowing they were facing not just heroic martyrdom, 0:16:00.200 --> 0:16:04.880 red cues, lack of medical care, splintered leadership, they still 0:16:04.880 --> 0:16:07.400 would have left the comfort of their home countries again 0:16:07.760 --> 0:16:12.160 places like America, England, Canada, to travel to Syria, an 0:16:12.160 --> 0:16:16.160 active war zone. So Jackson needed to understand what what 0:16:16.240 --> 0:16:18.680 ICE is doing to lure people in that was so 0:16:18.920 --> 0:16:21.760 very powerful. We identified the recruitment narratives and they were 0:16:21.920 --> 0:16:25.520 largely around people who thought that this was the devout, correct, 0:16:25.520 --> 0:16:28.400 religous thing to do, that this wasn't mslim utopia than 0:16:28.560 --> 0:16:31.000 was going to lead to a healthier, happier life than 0:16:31.160 --> 0:16:34.080 remaining in the West. People who are interested in the 0:16:34.080 --> 0:16:37.840 military conflated. So there were several and we generated a 0:16:37.920 --> 0:16:42.760 targeting strategy to reach these people based on their online browsing, 0:16:43.280 --> 0:16:47.080 specifically like their online searches take a moment to absorb that. 0:16:47.600 --> 0:16:51.240 Based on online browsing history and searches, Jigsaw were able 0:16:51.280 --> 0:16:55.120 to identify potential extremists and serve them adds to push 0:16:55.120 --> 0:16:59.240 them towards alternative content. But wouldn't that raise alarm bells 0:16:59.280 --> 0:17:02.320 for the people do in the searching Yasmin and her 0:17:02.360 --> 0:17:05.280 team had to be subtle. If you were interested in 0:17:05.359 --> 0:17:09.000 fact was about Jahard, religious udics about Johard, we would 0:17:09.000 --> 0:17:11.320 give you those originally, It's just not the ones that 0:17:11.359 --> 0:17:14.359 Isis was proposing, but alternative ones. Or if you're interested 0:17:14.359 --> 0:17:16.680 to understand what life is like in the Caliphate, let's 0:17:16.680 --> 0:17:20.200 show you citizen journalism of the long cueues for bread 0:17:20.440 --> 0:17:22.399 or the state of the hospitals in the Caliphate. So 0:17:22.400 --> 0:17:26.679 you're still getting something that addresses your interest. Um, it 0:17:26.680 --> 0:17:29.240 would just be an alternative information source to the one 0:17:29.280 --> 0:17:31.679 that you were looking for and it was really important 0:17:31.680 --> 0:17:34.040 to us to make sure that we were targeting, you know, 0:17:34.119 --> 0:17:36.240 and finding people who were really at risk as opposed 0:17:36.240 --> 0:17:38.760 to just people who are interested in isis that's right. 0:17:39.160 --> 0:17:42.560 Looking at browsing and search history, Jigsaw and Google can 0:17:42.600 --> 0:17:45.679 look at two different people looking up fatoas online and 0:17:45.760 --> 0:17:49.320 only serve redirecting ads to one. When you get the 0:17:49.480 --> 0:17:52.040 level more granular, you can start to set up a 0:17:52.080 --> 0:17:55.879 targeting strategy that really does differentiate mainstream interest in this 0:17:55.960 --> 0:17:59.320 group to the people who are you know, sympathetic and 0:17:59.520 --> 0:18:02.480 potential join us. It has turned out to be really 0:18:02.480 --> 0:18:07.280 effective since my conversation with Yasmin. A terrorist attack on 0:18:07.359 --> 0:18:11.879 two mosques in Christchurch, New Zealand killed fifty people and 0:18:11.960 --> 0:18:14.680 focus the world's attention on the threat posed by far 0:18:14.800 --> 0:18:17.960 right terror. I emailed with Yasmin and she told me 0:18:18.000 --> 0:18:20.359 that she and her team are quote attending to the 0:18:20.440 --> 0:18:26.520 far right threat with increased urgency. I think it's nice 0:18:26.560 --> 0:18:28.400 that they're trying to fix things, but you know, it's 0:18:28.400 --> 0:18:32.240 like the Anhauser Busch family funding a study on alcohol, right, 0:18:32.280 --> 0:18:34.760 I guess, except we kind of no alcohol is bad, 0:18:34.800 --> 0:18:38.000 whereas we can still hope that the internet is neutral, 0:18:38.400 --> 0:18:40.280 although it is clear the Internet has become a key 0:18:40.320 --> 0:18:43.399 method of radicalization for terrorists of all stripes, and so 0:18:43.720 --> 0:18:45.280 for me at least, it's good to know that the 0:18:45.320 --> 0:18:48.359 technology companies are acknowledging that and smart people like Yasmin 0:18:48.400 --> 0:18:51.520 are working on it. Although it does also mean living 0:18:51.520 --> 0:18:54.399 in a world where our main internet search providers also 0:18:54.560 --> 0:18:57.480 edit the results. Do we want Google controlling what we 0:18:57.560 --> 0:19:01.120 know about the world or recommending what else we should now? Right? 0:19:01.280 --> 0:19:03.280 And then what if Google turned that immense power to 0:19:03.320 --> 0:19:09.680 influence directly onto me or you? More on that after 0:19:09.720 --> 0:19:18.360 the break, Tristan, we've been talking about chicksaw on their 0:19:18.359 --> 0:19:22.399 manipulation when you look at elections and the way extremist 0:19:22.480 --> 0:19:27.760 groups are using technology to radicalize and proliferate. Should tech 0:19:27.800 --> 0:19:30.240 companies be trying to offset some of the damage they're 0:19:30.240 --> 0:19:33.720 doing by intervening for good? Yeah, Well, this is an 0:19:33.720 --> 0:19:38.200 incredibly nuanced topic because essentially the way to see what's 0:19:38.200 --> 0:19:41.399 going on with technology today is there's a massive asymmetry. 0:19:41.600 --> 0:19:43.040 It would be easy for you and I and he 0:19:43.040 --> 0:19:46.560 has mean to say clearly isis is bad and terrorism 0:19:46.640 --> 0:19:50.400 is bad. So let's redirect when you're searching for things 0:19:50.400 --> 0:19:53.760 that look like terrorist videos. But if you suddenly then 0:19:53.760 --> 0:19:56.639 open the door and say you're about to watch a 0:19:56.760 --> 0:20:01.320 video on climate change, and people at Google, for whatever reason, 0:20:01.359 --> 0:20:04.040 believe that climate change wasn't real and they said, we're 0:20:04.040 --> 0:20:07.159 going to start redirecting you away from that. Again, who 0:20:07.200 --> 0:20:09.880 would they be to say? And that's actually a critical 0:20:09.960 --> 0:20:12.440 juncture that we're at right now, because there's kind of 0:20:12.520 --> 0:20:15.560 a pluralism where we all believe and think different things, 0:20:15.560 --> 0:20:17.119 and we want to say that's your truth and not 0:20:17.200 --> 0:20:19.200 my truth, and so we have this collapse of well, 0:20:19.240 --> 0:20:23.280 who is the moral authority on these topics? And after all, 0:20:23.640 --> 0:20:27.399 we're all Americans, well not me, but a citizens of 0:20:27.400 --> 0:20:30.919 a democracy. We have a voice in our diplomacy. What 0:20:31.000 --> 0:20:33.800 about is uses as the Internet? I think it's important 0:20:33.800 --> 0:20:36.680 that we not try to just have a philosophy cocktail 0:20:36.720 --> 0:20:39.840 conversation and debate it. We have to actually recognize that 0:20:39.880 --> 0:20:43.120 this is having real world consequences. But the difficult thing 0:20:43.200 --> 0:20:46.119 to reckon with is that it all comes back to attention. 0:20:46.640 --> 0:20:48.639 YouTube has a tilt. We're on the one side of 0:20:48.640 --> 0:20:54.520 the spectrum you have the calm science Carl Siggin slowly explaining, 0:20:54.640 --> 0:20:57.440 and on the other side of the spectrum. You have UFOs, Bigfoot, 0:20:57.440 --> 0:21:01.240 conspiracy theories, and crazy Town. If I'm YouTube and I 0:21:01.280 --> 0:21:03.399 want you to watch more, Let's say you start at 0:21:03.400 --> 0:21:05.400 the science section. If I want you to watch more, 0:21:05.600 --> 0:21:07.960 which direction between those two am I going to steer you? 0:21:08.520 --> 0:21:11.560 I'm always going to steer you too. Crazy Town so 0:21:11.800 --> 0:21:14.639 at scale, even in language that the engineers don't speak, 0:21:14.840 --> 0:21:17.880 the algorithms had figured out that crazy Town is really 0:21:17.920 --> 0:21:20.080 really good at getting people's attention. And we know the 0:21:20.119 --> 0:21:23.760 effect isn't just on politics, what about regular people who 0:21:23.760 --> 0:21:27.200 get pushed into joinning ISIS or taking other extreme positions. 0:21:27.480 --> 0:21:29.840 If you start a team girl on a dieting video, 0:21:30.080 --> 0:21:34.080 the YouTube algorithm recommends an arexia videos because it's calculating 0:21:34.080 --> 0:21:36.000 away and figures out those are the things that are 0:21:36.000 --> 0:21:39.840 really good at keeping people of that age demographic on YouTube. 0:21:39.920 --> 0:21:42.280 And if you start someone on a Supember eleventh news 0:21:42.359 --> 0:21:45.760 video of the planes crashing into the towers, the recommendations 0:21:45.760 --> 0:21:48.280 are all going to be the nine eleven conspiracy theory videos. 0:21:48.280 --> 0:21:53.679 And Alex Jones YouTube actually recommended fifteen billion times Alex 0:21:53.760 --> 0:21:57.080 Jones videos, two billion of which were viewed, even if 0:21:57.119 --> 0:21:59.600 only one out of a thousand of those people believed it, 0:21:59.680 --> 0:22:01.880 So one out of a thousand people watching those two 0:22:01.880 --> 0:22:04.520 billion views believe that that's like printing, you know, a 0:22:04.600 --> 0:22:07.560 church of scientology called about you know once a month. 0:22:07.600 --> 0:22:11.160 In terms of the scale, we are jacked into these systems, 0:22:11.160 --> 0:22:14.640 We are jacked into these environments that are telling us 0:22:14.680 --> 0:22:17.600 and steering us towards flows of attention and ways of 0:22:17.640 --> 0:22:20.840 seeing the world that are inevitable. And the question is 0:22:20.960 --> 0:22:24.679 what flows of attention do we want? And in the end, 0:22:24.720 --> 0:22:27.480 it all comes down to who's steering the ship and 0:22:27.520 --> 0:22:31.879 what they're steering you towards. Who watches the watchman he 0:22:32.000 --> 0:22:35.400 has been argues that jakes Will takes ethics heavily into consideration, 0:22:36.240 --> 0:22:39.560 but not all companies do. And Tristan actually has a 0:22:39.640 --> 0:22:43.280 radical idea. He believes that we need to wean ourselves 0:22:43.480 --> 0:22:47.000 as an economy and as a society from business models 0:22:47.040 --> 0:22:49.560 that compete for our attention to sell us things. I 0:22:49.560 --> 0:22:53.200 think we really underestimate that this is affecting every layer 0:22:53.320 --> 0:22:56.200 of society. And when you have three or four tech 0:22:56.240 --> 0:22:59.560 companies choosing what will show up in the minds of 0:22:59.560 --> 0:23:02.119 two bill and people every day, you know, our minds 0:23:02.160 --> 0:23:04.920