1 00:00:14,702 --> 00:00:22,030 Herald (H): Yeah. Welcome to our next talk, Social Cooling. You know, people say 2 00:00:22,030 --> 00:00:28,696 "I have no problem with surveillance. I have nothing to hide," but then, you know, 3 00:00:28,696 --> 00:00:37,453 maybe the neighbors and maybe this and maybe that. So, tonight we're going to 4 00:00:37,453 --> 00:00:45,210 hear Tijmen Schep who's from Holland. He's a privacy designer and a freelance 5 00:00:45,210 --> 00:00:53,193 security researcher and he's gonna hold a talk about how digital surveillance 6 00:00:53,193 --> 00:01:04,179 changes our social way of interacting. So, please, let's have a hand for Tijmen Schep! 7 00:01:04,179 --> 00:01:14,710 *Applause* 8 00:01:14,710 --> 00:01:18,289 Tijmen Schep (TS): Hi everyone. Really cool that you're all here and really happy 9 00:01:18,289 --> 00:01:24,430 to talk here. It's really an honor. My name is Tijmen Schep and I am a technology 10 00:01:24,430 --> 00:01:30,889 critic. And that means that it's my job to not believe [audio cuts out] tells us and 11 00:01:30,889 --> 00:01:36,810 that's really a lot of fun. [audio cuts out] is, how do I get a wider audience 12 00:01:36,810 --> 00:01:40,350 involved in understanding technology and the issues that are arising from 13 00:01:40,350 --> 00:01:46,049 technology? Because I believe that change comes when the public demands it. I think 14 00:01:46,049 --> 00:01:51,299 that's really one of the important things when change happens. And for me as a 15 00:01:51,299 --> 00:01:55,279 technology critic, for me words are very much how I hack the system, how I try to 16 00:01:55,279 --> 00:02:01,579 hack this world. And so, tonight I'm going to talk to you about one of these words 17 00:02:01,579 --> 00:02:08,310 that I think could help us. Framing the issue is half the battle. [audio cuts out] 18 00:02:08,310 --> 00:02:13,069 and frame the problem - if we can explain, what the problem is in a certain frame... 19 00:02:13,069 --> 00:02:17,990 that, you know, makes certain positions already visible, that's really half the 20 00:02:17,990 --> 00:02:24,691 battle won. So, that frame is social cooling. But before I go into it, I want 21 00:02:24,691 --> 00:02:31,880 to ask you a question. Who here recognizes this? You're on Facebook or some other 22 00:02:31,880 --> 00:02:38,990 social site, and you click on the link because you think "Oh I could [audio cuts 23 00:02:38,990 --> 00:02:43,650 out] listen [audio cuts out] could click on this, but it might look bad. It might 24 00:02:43,650 --> 00:02:46,501 be remembered by someone. Some agency might remember it, and I could click on 25 00:02:46,501 --> 00:02:51,960 it, but I'm hesitating to click." *Microphone buzzing* 26 00:02:51,960 --> 00:02:54,960 *laughter* 27 00:03:02,380 --> 00:03:07,340 TS: That better? Can everyone hear me now? Audience: No. 28 00:03:07,340 --> 00:03:16,570 TS: No. Okay, that... yeah. Should I start again? Okay. So, you're on Facebook, and 29 00:03:16,570 --> 00:03:19,850 you're thinking "Oh, that's an interesting link. I could click on that," but you're 30 00:03:19,850 --> 00:03:22,600 hesitating because maybe someone's gonna remember 31 00:03:22,600 --> 00:03:26,320 that. And that might come back to me later, and who here recognizes that 32 00:03:26,320 --> 00:03:31,581 feeling? So, pretty much almost everybody. And that's increasingly what I find, when 33 00:03:31,581 --> 00:03:36,520 I talk about the issue, that people really start to recognize this. And I think a 34 00:03:36,520 --> 00:03:41,540 word we could use to describe that is "Click Fear." This hesitation, it could be 35 00:03:41,540 --> 00:03:46,261 click fear. And you're not alone. Increasingly, we find that, research 36 00:03:46,261 --> 00:03:50,880 points that this is a wide problem, that people are hesitating to click some of the 37 00:03:50,880 --> 00:03:55,090 links. For example, after the Snowden revelations, people were less likely to 38 00:03:55,090 --> 00:03:59,290 research issues about terrorism and other things on Wikipedia because they thought 39 00:03:59,290 --> 00:04:03,520 "Well, maybe the NSA wouldn't like it if I [audio cuts out] that. Okay, not gonna 40 00:04:03,520 --> 00:04:10,780 move. And visiting Google as well. So this is a pattern that there's research... are 41 00:04:10,780 --> 00:04:15,380 pointing to. And it's not very strange, of course. I mean, we all understand that if 42 00:04:15,380 --> 00:04:17,661 you feel you're being watched, you change your behavior. It's a 43 00:04:17,661 --> 00:04:23,400 very logical thing that we all understand. And I believe that technology is really 44 00:04:23,400 --> 00:04:27,440 amplifying this effect. I think that's something that we really have to come to 45 00:04:27,440 --> 00:04:32,090 grips with. And that's why I think social cooling could be useful with that. Social 46 00:04:32,090 --> 00:04:36,249 cooling describes in a way how in increasingly digital world, where our 47 00:04:36,249 --> 00:04:42,870 digital lives are increasingly digitized, it becomes easier to feel this pressure, to 48 00:04:42,870 --> 00:04:49,220 feel these normative effects of these systems. And very much you see that, 49 00:04:49,220 --> 00:04:52,740 because increasingly, your data is being turned into thousands of scores by data 50 00:04:52,740 --> 00:04:56,289 brokers and other companies. And those scores are increasing influences you're... 51 00:04:56,289 --> 00:05:01,349 influencing your chances in life. And this is creating an engine of oppression, an 52 00:05:01,349 --> 00:05:07,990 engine of change that we have to understand. And the fun thing is that in a 53 00:05:07,990 --> 00:05:13,689 way this idea is really being helped by Silicon Valley, who for a long time has 54 00:05:13,689 --> 00:05:17,120 said "Data is the new gold," but they've recently, in the last five years, changed 55 00:05:17,120 --> 00:05:22,029 that narrative. Now they're saying "Data is the new oil," and that's really funny, 56 00:05:22,029 --> 00:05:25,270 because if data is the new oil, then immediately you get the question "Wait, 57 00:05:25,270 --> 00:05:31,210 oil gave us global warming, so then, what does data give us?" And I believe that if 58 00:05:31,210 --> 00:05:35,319 oil leads to global warming, then data could lead to social cooling. That could 59 00:05:35,319 --> 00:05:40,360 be the word that we use for these negative effects of big data. In order to really 60 00:05:40,360 --> 00:05:43,169 understand this, and go into it, we have to look at three things. First, we're 61 00:05:43,169 --> 00:05:47,090 going to talk about the reputation economy, how that system works. Second 62 00:05:47,090 --> 00:05:51,099 chapter, we're going to look at behavior change, how it is influencing us and 63 00:05:51,099 --> 00:05:55,419 changing our behavior. And finally, to not let you go home depressed, I'm gonna talk 64 00:05:55,419 --> 00:06:01,939 about how can we deal with this. So first. The reputation economy. Already we've seen 65 00:06:01,939 --> 00:06:08,152 today that China is building this new system, the social credit system. It's a 66 00:06:08,152 --> 00:06:11,210 system that will give every citizen in China a score that basically represents 67 00:06:11,210 --> 00:06:15,650 how well-behaved they are. And it will influence your ability to get a job, 68 00:06:15,650 --> 00:06:21,349 a loan, a visum and even a date. And for example, the current version of the 69 00:06:21,349 --> 00:06:26,339 system, Sesame Credit, one of the early prototypes, already gives everybody that 70 00:06:26,339 --> 00:06:30,689 wants to a score, but it also is connected to the largest dating website in China. 71 00:06:30,689 --> 00:06:34,810 So, you can kind of find out "Is this person that I'm dating... what kind of 72 00:06:34,810 --> 00:06:40,819 person is this? Is this something who's, you know, well viewed by Chinese society?" 73 00:06:42,239 --> 00:06:45,189 This is where it gets really heinous for me, because until now you could say "Well, 74 00:06:45,189 --> 00:06:48,860 these reputation systems, they're fair, if you're a good person, you get a higher 75 00:06:48,860 --> 00:06:52,470 score. If you're bad person, you get a lower score," but it's not that simple. I 76 00:06:52,470 --> 00:06:55,499 mean, your friends' score influences your score, and your score influences your 77 00:06:55,499 --> 00:07:00,759 friends' score, and that's where you really start to see how complex social 78 00:07:00,759 --> 00:07:04,581 pressures arrive, and where we can see the effects of data stratification, where 79 00:07:04,581 --> 00:07:07,389 people are starting to think "Hey, who are my friends, and who should I be friends 80 00:07:07,389 --> 00:07:14,350 with?" You could think "That only happens in China. Those Chinese people are, you 81 00:07:14,350 --> 00:07:18,999 know, different." But the exact same thing is happening here in the West, 82 00:07:18,999 --> 00:07:22,031 except we're letting the market build it. I'll give you an example. This is a 83 00:07:22,031 --> 00:07:25,750 company called "deemly" - a Danish company - and this is their video for their 84 00:07:25,750 --> 00:07:28,849 service. Video narrator (VN): ... renting 85 00:07:28,849 --> 00:07:33,600 apartments from others, and she loves to swap trendy clothes and dresses. She's 86 00:07:33,600 --> 00:07:38,059 looking to capture her first lift from a RideShare app, but has no previous reviews 87 00:07:38,059 --> 00:07:40,020 to help support her. Video background voices: Awww. 88 00:07:40,020 --> 00:07:44,259 VN: Luckily, she's just joined deemly, where her positive feedback from the other 89 00:07:44,259 --> 00:07:50,699 sites appears as a deemly score, helping her to win a RideShare in no time. Deemly 90 00:07:50,699 --> 00:07:55,879 is free to join and supports users across many platforms, helping you to share and 91 00:07:55,879 --> 00:08:01,159 benefit from the great reputation you've earned. Imagine the power of using your 92 00:08:01,159 --> 00:08:04,040 deemly score alongside your CV for a job application... 93 00:08:04,040 --> 00:08:06,159 TS: Like in China. VN: ... perhaps to help get a bank loan... 94 00:08:06,159 --> 00:08:08,339 TS: Like... VN: or even to link to from your dating 95 00:08:08,339 --> 00:08:09,339 profile. TS: Like in China! 96 00:08:09,339 --> 00:08:15,759 VN: Sign up now at deemly.co. Deemly: better your sharing. 97 00:08:15,759 --> 00:08:21,590 *Applause* TS: Thanks. There is a change. There is 98 00:08:21,590 --> 00:08:26,499 difference, though. The funny thing about here is that it's highly invisible to us. 99 00:08:26,499 --> 00:08:29,580 The Chinese government is very open about what they're building, but here we are 100 00:08:29,580 --> 00:08:33,360 very blind to what's going on. Mostly, when we talk about these things, then 101 00:08:33,360 --> 00:08:37,450 we're talking about these systems that give us a very clear rating, like Airbnb, 102 00:08:37,450 --> 00:08:41,909 Uber, and of course the Chinese system. The thing is, most of these systems are 103 00:08:41,909 --> 00:08:47,350 invisible to us. There's a huge market of data brokers who are, you know, not 104 00:08:47,350 --> 00:08:52,330 visible to you, because you are not the customer. You are the product. And these 105 00:08:52,330 --> 00:08:57,160 data brokers, well, what they do is, they gather as much data as possible about you. 106 00:08:57,160 --> 00:09:03,590 And that's not all. They then create up to eight thousand scores about you. In the 107 00:09:03,590 --> 00:09:07,820 United States, these companies have up to 8,000 scores, and in Europe it's a little 108 00:09:07,820 --> 00:09:13,190 less, around 600. These are scores about things like your IQ, your psychological 109 00:09:13,190 --> 00:09:19,020 profile, your gullibility, your religion, your estimated life span. 8,000 of these 110 00:09:19,020 --> 00:09:23,690 different things about you. And how does that work? Well, it works by machine 111 00:09:23,690 --> 00:09:28,580 learning. So, machine learning algorithms can find patterns in society that we can 112 00:09:28,580 --> 00:09:35,750 really not anticipate. For example, let's say you're a diabetic, and, well, let's 113 00:09:35,750 --> 00:09:40,880 say this data broker company has a mailing list, or has an app, that diabetic 114 00:09:40,880 --> 00:09:44,311 patients use. And they also have the data of these diabetic patients about what they 115 00:09:44,311 --> 00:09:48,440 do on Facebook. Well, there you can start to see correlations. So, if diabetic 116 00:09:48,440 --> 00:09:53,840 patients more often like gangster-rap and pottery on Facebook, well, then you could 117 00:09:53,840 --> 00:09:58,120 deduce from that if you also like gangster-rap or pottery on Facebook, then 118 00:09:58,120 --> 00:10:03,460 perhaps you also are more likely to have or get diabetes. It is highly 119 00:10:03,460 --> 00:10:08,690 unscientific, but this is how the system works. And this is an example of how that 120 00:10:08,690 --> 00:10:13,470 works with just your Facebook scores. Woman in the video: ... see was lowest about 121 00:10:13,470 --> 00:10:17,570 60% when it came to predicting whether a user's parents were still together when 122 00:10:17,570 --> 00:10:21,310 they were 21. People whose parents divorced before they were 21 tended to 123 00:10:21,310 --> 00:10:26,330 like statements about relationships. Drug users were ID'd with about 65% accuracy. 124 00:10:26,330 --> 00:10:33,120 Smokers with 73%, and drinkers with 70%. Sexual orientation was also easier to 125 00:10:33,120 --> 00:10:40,400 distinguish among men. 88% right there. For women, it was about 75%. Gender, by 126 00:10:40,400 --> 00:10:44,870 the way, race, religion, and political views, were predicted with high accuracy 127 00:10:44,870 --> 00:10:50,270 as well. For instance: White versus black: 95%. 128 00:10:50,270 --> 00:10:54,070 TS: So, the important thing to understand here is that this isn't really about your 129 00:10:54,070 --> 00:10:57,890 data anymore. Like, oftentimes when we talk about data protection, we talk about 130 00:10:57,890 --> 00:11:03,190 "Oh, I want to keep control of my data." But this is their data. This data that 131 00:11:03,190 --> 00:11:10,190 they deduce, that they derive from your data. These are opinions about you. And 132 00:11:10,190 --> 00:11:14,300 these things are what, you know, make it so that even though you never filled in a 133 00:11:14,300 --> 00:11:18,960 psychological test, they'd have one. A great example of that, how that's used, is 134 00:11:18,960 --> 00:11:24,220 a company called Cambridge Analytica. This company has created detailed profiles 135 00:11:24,220 --> 00:11:28,690 about us through what they call psychographics and I'll let them explain 136 00:11:28,690 --> 00:11:33,100 it themselves. Man in the video: By having hundreds and 137 00:11:33,100 --> 00:11:36,930 hundreds of thousands of Americans undertake this survey, we were able to 138 00:11:36,930 --> 00:11:40,090 form a model to predict the personality of every 139 00:11:40,090 --> 00:11:46,230 single adult in the United States of America. If you know the personality of 140 00:11:46,230 --> 00:11:50,300 the people you're targeting you can nuance your messaging to resonate more 141 00:11:50,300 --> 00:11:55,640 effectively with those key audience groups. So, for a highly neurotic and 142 00:11:55,640 --> 00:12:01,200 conscientious audience, you're going to need a message that is rational and fair- 143 00:12:01,200 --> 00:12:05,700 based, or emotionally-based. In this case, the threat of a burglary, and the 144 00:12:05,700 --> 00:12:10,410 insurance policy of a gun is very persuasive. And we can see where these 145 00:12:10,410 --> 00:12:15,280 people are on the map. If we wanted to drill down further, we could resolve the 146 00:12:15,280 --> 00:12:19,780 data to an individual level, where we have somewhere close to four or five thousand 147 00:12:19,780 --> 00:12:24,460 data points on every adult in the United States. 148 00:12:24,460 --> 00:12:28,120 TS: So, yeah. This is the company that worked with both the... for the Brexit 149 00:12:28,120 --> 00:12:33,570 campaign and with the Trump campaign. Of course, little after Trump campaign, all 150 00:12:33,570 --> 00:12:37,510 the data was leaked, so data on 200 million Americans was leaked, And 151 00:12:37,510 --> 00:12:40,380 increasingly, you can see this data described as "modeled voter 152 00:12:40,380 --> 00:12:46,440 ethnicities and religions." So, this is this derived data. You might think that 153 00:12:46,440 --> 00:12:49,630 when you go online and use Facebook and use all these services, that advertisers 154 00:12:49,630 --> 00:12:53,310 are paying for you. That's a common misperception. That's not really the case. 155 00:12:53,310 --> 00:12:57,500 What's really going on is that, according to SSC research, the majority of the 156 00:12:57,500 --> 00:13:01,980 money made in this data broker market is made from risk management. All right, so, 157 00:13:01,980 --> 00:13:07,380 in a way you could say that it's not really marketers that are paying for you, 158 00:13:07,380 --> 00:13:12,930 it's your bank. It's ensurers. It's your employer. It's governments. These kind of 159 00:13:12,930 --> 00:13:19,210 organizations are the ones who buy these profiles. The most. More than the other 160 00:13:19,210 --> 00:13:23,530 ones. Of course, the promise of big data is that you can then manage risk. Big data 161 00:13:23,530 --> 00:13:27,720 is the idea that with data you can understand things and then manage them. 162 00:13:27,720 --> 00:13:31,700 So what really is innovation in this big data world, this data economy, is the 163 00:13:31,700 --> 00:13:36,110 democratization of the background check. That's really the core of this, this market 164 00:13:36,110 --> 00:13:39,310 that now you can find out everything about everyone. 165 00:13:39,310 --> 00:13:44,810 So, yeah, now your... in past, only perhaps your bank could know your credit score but 166 00:13:44,810 --> 00:13:47,980 now your green grocer knows your psychological profile. 167 00:13:47,980 --> 00:13:55,070 Right that's a new level of, yeah, what's going on here. It's not only inv... not 168 00:13:55,070 --> 00:13:58,880 only invisible but it's also huge according to the same research by the FCC 169 00:13:58,880 --> 00:14:05,340 this market was already worth 150 billion dollars in 2015. So, it's invisible, it's 170 00:14:05,340 --> 00:14:10,760 huge and hardly anyone knows about it. But that's probably going to change. And that 171 00:14:10,760 --> 00:14:18,040 brings us to the second part: Behavioral change. We already see this first part of 172 00:14:18,040 --> 00:14:21,370 this, how behavioral change is happening through these systems. That's through outside 173 00:14:21,370 --> 00:14:25,520 influence and we've, we've talked a lot about this in this conference. For example 174 00:14:25,520 --> 00:14:29,110 we see how Facebook and advertisers try to do that. We've also seen how China is 175 00:14:29,110 --> 00:14:32,870 doing that, trying to influence you. Russia has recently tried to use Facebook to 176 00:14:32,870 --> 00:14:36,710 influence the elections and of course companies like Cambridge Analytica try to 177 00:14:36,710 --> 00:14:39,420 do the same thing. And here you can have a debate on, you 178 00:14:39,420 --> 00:14:43,480 know, to what extent are they really influencing us, but I think that's not 179 00:14:43,480 --> 00:14:48,090 actually the really, the most interesting question. What interests me most of all is 180 00:14:48,090 --> 00:14:53,840 how we are doing it ourselves, how we are creating new forms of self-censorship and 181 00:14:53,840 --> 00:14:59,040 and are proactively anticipating these systems. Because once you realize that 182 00:14:59,040 --> 00:15:01,870 this is really about risk management you start... and this is about banks and 183 00:15:01,870 --> 00:15:06,400 employers trying to understand you, people start to understand that this will go beyond 184 00:15:06,400 --> 00:15:10,470 click fear, if you remember. This will go beyond, this will become, you know, 185 00:15:10,470 --> 00:15:14,230 when people find out this will be, you know, not getting a job for example. 186 00:15:14,230 --> 00:15:18,980 This'll be about getting really expensive insurance. It'll be about all these kinds 187 00:15:18,980 --> 00:15:22,080 of problems and people are increasingly finding this out. So for example in the 188 00:15:22,080 --> 00:15:28,970 United States if you... the IRS might now use data profile... are now using data 189 00:15:28,970 --> 00:15:33,860 profiles to find out who they should audit. So I was talking recently to a girl 190 00:15:33,860 --> 00:15:37,560 and and she said: "Oh I recently tweeted about... a negative tweet about the IRS," and 191 00:15:37,560 --> 00:15:39,800 she immediately grabbed her phone to delete it. 192 00:15:39,800 --> 00:15:44,600 When she realized that, you know, this could now be used against her in a way. 193 00:15:44,600 --> 00:15:49,320 And that's the problem. Of course we see all kinds of other crazy examples that the big... 194 00:15:49,320 --> 00:15:54,450 the audience that we measure... the wider public is picking up on, like who... so we now have 195 00:15:54,450 --> 00:15:58,610 algorithms that can find out if you're gay or not. And these things scare people and 196 00:15:58,610 --> 00:16:03,460 these things are something we have to understand. So, chilling effects this what 197 00:16:03,460 --> 00:16:08,380 this boils down to. For me, more importantly than these influences of these big 198 00:16:08,380 --> 00:16:12,850 companies and nation states is how people themselves are experiencing these chilling 199 00:16:12,850 --> 00:16:17,860 effects like you yourself have as well. That brings us back to social cooling. For 200 00:16:17,860 --> 00:16:23,810 me, social cooling is about these two things combined at once and this 201 00:16:23,810 --> 00:16:28,720 increasing ability of agents and... and groups to influence you and on the other hand the 202 00:16:28,720 --> 00:16:33,240 increasing willingness of people themselves to change their own behavior to 203 00:16:33,240 --> 00:16:38,680 proactively engage with this issue. There are three long-term consequences 204 00:16:38,680 --> 00:16:44,120 that I want to dive into. The first is how this affects the individual, the second is 205 00:16:44,120 --> 00:16:49,320 how it affects society, and the third is how it affects the market. So let's look 206 00:16:49,320 --> 00:16:55,350 the individual. Here we've seen, there's a rising culture of self-censorship. It 207 00:16:55,350 --> 00:16:58,970 started for me with an article that I read in New York Times, where a student was saying: 208 00:16:58,970 --> 00:17:02,430 "Well we're very very reserved." She's going to do things like spring break. 209 00:17:02,430 --> 00:17:05,819 I said: "Well you don't have to defend yourself later," so you don't do it. 210 00:17:05,819 --> 00:17:08,859 And what she's talking about, she's talking about doing crazy things, you 211 00:17:08,859 --> 00:17:12,410 know, letting go, having fun. She's worried that the next day it'll be on 212 00:17:12,410 --> 00:17:15,839 Facebook. So what's happening here is that you do 213 00:17:15,839 --> 00:17:18,190 have all kinds of freedoms: You have the freedom to look up things, you have the 214 00:17:18,190 --> 00:17:22,290 freedom to to say things, but you're hesitating to use it. And that's really 215 00:17:22,290 --> 00:17:28,460 insidious. That has an effect on a wider society and here we really see the 216 00:17:28,460 --> 00:17:34,260 societal value of privacy. Because in society often minority values later become 217 00:17:34,260 --> 00:17:40,559 majority values. An example is... is weed. I'm from... I'm from the Netherlands and 218 00:17:40,559 --> 00:17:45,059 there you see, you know, at first it's something that you just don't do and it's 219 00:17:45,059 --> 00:17:48,800 you know a bit of a "uhh", but then "Oh, maybe yeah, you should... you should try it as well," and 220 00:17:48,800 --> 00:17:53,020 people try it and slowly under the surface of the society, people change their minds 221 00:17:53,020 --> 00:17:55,980 about these things. And then, after a while it's like, you know, "What are we still 222 00:17:55,980 --> 00:17:59,300 worried about?" How the same pattern help it happens of 223 00:17:59,300 --> 00:18:03,070 course with way bigger things like this: Martin Luther King: "I must honestly say to 224 00:18:03,070 --> 00:18:10,470 you that I never intend to adjust myself to racial segregation and discrimination." 225 00:18:12,410 --> 00:18:14,100 TS: This is the same pattern that's happening 226 00:18:14,100 --> 00:18:17,680 for all kinds of things that that change in society, and that's what privacy is so 227 00:18:17,680 --> 00:18:20,370 important for, and that's why it's so important that people have the ability to 228 00:18:20,370 --> 00:18:22,880 look things up and to change their minds and to talk about each other without 229 00:18:22,880 --> 00:18:27,930 feeling so watched all the time. The third thing is how this impacts the 230 00:18:27,930 --> 00:18:33,770 market. Here we see very much the rise of a culture of risk avoidance. An example 231 00:18:33,770 --> 00:18:36,580 here is that in 1995 already, doctors in New York were 232 00:18:36,580 --> 00:18:43,510 given scores, and what happened was that the doctors who try to help advanced 233 00:18:43,510 --> 00:18:47,010 stage cancer patients, complex patients, who try to do the operation, difficult 234 00:18:47,010 --> 00:18:52,380 operations, got a low score, because these people more often died, while doctors that 235 00:18:52,380 --> 00:18:56,800 didn't lift a finger and didn't try to help got a high score. Because, well, people 236 00:18:56,800 --> 00:19:00,780 didn't die. So you see here that these systems that, they bring all kinds of 237 00:19:00,780 --> 00:19:04,670 perverse incentives. They, you know, they're they lower the willingness for everybody 238 00:19:04,670 --> 00:19:07,760 to take a risk and in some areas of society we really like people to take 239 00:19:07,760 --> 00:19:15,150 risks. They're like entrepreneurs, doctors. So in the whole part you could say that 240 00:19:15,150 --> 00:19:19,010 this, what we're seeing here, is some kind of trickle-down risk aversion, where 241 00:19:19,010 --> 00:19:23,130 the willing, the... the way that companies and governments want to 242 00:19:23,130 --> 00:19:27,440 manage risk, that's trickling down to us. And we're we of course want them to like 243 00:19:27,440 --> 00:19:30,309 us, want to have a job, we want to have insurance, and then we increasingly start 244 00:19:30,309 --> 00:19:36,830 to think "Oh, maybe I should not do this." It's a subtle effect. So how do we deal with 245 00:19:36,830 --> 00:19:39,960 this? Well, together. I think this is a really 246 00:19:39,960 --> 00:19:43,490 big problem. I think this is such a big problem that, that it can't be managed by 247 00:19:43,490 --> 00:19:47,460 just some, some hackers or nerds, building something, or by politicians, making a law. 248 00:19:47,460 --> 00:19:52,640 This is a really a society-wide problem. So I want to talk about all these groups 249 00:19:52,640 --> 00:19:57,920 that should get into this: the public, politicians, business, and us. 250 00:19:57,920 --> 00:20:03,270 So the public. I think we have to talk about and maybe extend the metaphor of the 251 00:20:03,270 --> 00:20:06,890 cloud and say we have to learn to see the stars behind the cloud. Alright, that's one 252 00:20:06,890 --> 00:20:11,679 way that we could... that's a narrative we could use. I really like to use humor to 253 00:20:11,679 --> 00:20:17,050 explain this to a wider audience, so for example, last year I was part of an 254 00:20:17,050 --> 00:20:22,240 exhibits... helped develop exhibits about dubious devices and one of the devices 255 00:20:22,240 --> 00:20:25,010 there was called "Taste your status" which was a coffee machine that gave you 256 00:20:25,010 --> 00:20:30,210 coffee based on your area code. So if you live in a good area code, you get nice coffee. 257 00:20:30,210 --> 00:20:33,630 You live in a bad area code, you get bad coffee. 258 00:20:33,630 --> 00:20:36,960 *music* *laugher* 259 00:20:36,960 --> 00:20:41,820 *applause* I'll go into it but... these... often times 260 00:20:41,820 --> 00:20:44,130 you can use humor to explain these things to a wider audience. I really like that 261 00:20:44,130 --> 00:20:47,670 method, that approach. We've got a long way to go though. I mean, 262 00:20:47,670 --> 00:20:50,150 if we look at the long, you know, how long it took for us to understand global 263 00:20:50,150 --> 00:20:53,150 warming, to really, you know, come to a stage where most people understand what it is 264 00:20:53,150 --> 00:20:58,260 and care about it except Donald Trump. Well, with data we really got a long way to 265 00:20:58,260 --> 00:21:01,890 go, we're really at the beginning of understanding this issue like this. 266 00:21:01,890 --> 00:21:07,970 Okay, so the second group that has to really wake up is politicians. And they have 267 00:21:07,970 --> 00:21:10,910 to understand that this is really about the balance of power. This is really about 268 00:21:10,910 --> 00:21:16,380 power. And if you permit me, I'll go into the big picture a little bit, as a media 269 00:21:16,380 --> 00:21:21,750 theorist. So this is called Giles Deleuze. he's a French philosopher and he explained 270 00:21:21,750 --> 00:21:26,000 in his work something that I find really useful, He said you have two systems of 271 00:21:26,000 --> 00:21:30,152 control in society and the one is the institutional one and that's the one we 272 00:21:30,152 --> 00:21:32,940 all know. You know that the judicial system so 273 00:21:32,940 --> 00:21:38,140 you're free to do what you want but then you cross a line you cross a law and the 274 00:21:38,140 --> 00:21:40,610 police get you you go for every charge you go to prison. That's the system we 275 00:21:40,610 --> 00:21:43,010 understand. But he says there's another system which 276 00:21:43,010 --> 00:21:46,840 is the social system this is a social pressure system and this for a long time 277 00:21:46,850 --> 00:21:49,670 wasn't really designed. But now increasingly we are able to do 278 00:21:49,670 --> 00:21:53,480 that so this is the system where you perform suboptimal behavior and then that 279 00:21:53,480 --> 00:21:58,240 gets measured and judged and then you get subtly nudged in the right direction. And 280 00:21:58,240 --> 00:22:01,130 there's some very important differences between these 2 systems. The institutional 281 00:22:01,130 --> 00:22:05,010 system you know it has this idea that you're a free citizen that makes up your 282 00:22:05,010 --> 00:22:10,500 own mind and you know what social system is like that's working all the time, 283 00:22:10,500 --> 00:22:13,280 constantly it doesn't matter if you're guilty or innocent it's always trying to 284 00:22:13,280 --> 00:22:19,020 push you. The old system, the institutional system is very much about punishment so if 285 00:22:19,020 --> 00:22:21,020 you break the rules you get punishment but 286 00:22:21,020 --> 00:22:23,570 people sometimes don't really care about punishment sometimes it's cool to get 287 00:22:23,570 --> 00:22:27,670 punishment. But the social system uses something way more powerful which is the 288 00:22:27,670 --> 00:22:33,890 fear of exclusion. We are social animals and we really care to belong to a group. 289 00:22:33,890 --> 00:22:37,350 The other difference is that it's very important that the institutional system is 290 00:22:37,350 --> 00:22:40,920 accountable. You know democratically to us how the social system at the moment is 291 00:22:40,920 --> 00:22:45,220 really really invisible like these algorithms how they work where the data is 292 00:22:45,220 --> 00:22:48,720 going it's very hard to understand and of course it's exactly what China loved so 293 00:22:48,720 --> 00:22:52,530 much about it right there's no you can stand in front of a tank but you can't 294 00:22:52,530 --> 00:22:56,690 really stand in front of the cloud. So yeah that's that's great it also helps me 295 00:22:56,690 --> 00:23:00,660 to understand when people say I have nothing to hide. I really understand that 296 00:23:00,660 --> 00:23:03,410 because when people say I have nothing to hide what they're saying is I have nothing 297 00:23:03,410 --> 00:23:06,340 to hide from the old system from the classic system from the institutional 298 00:23:06,340 --> 00:23:09,460 system. They're saying I want to help the police I trust our gover 299 00:23:09,460 --> 00:23:13,382 nment I trust our institutions and that's actually really a positive thing to say. 300 00:23:13,382 --> 00:23:16,980 The thing is they don't really see the other part of the system how increasingly there 301 00:23:16,980 --> 00:23:21,970 are parts that are not in your control they're not democratically checked and that's 302 00:23:21,970 --> 00:23:28,499 really a problem. So the third thing that I think we have to wake up is business, 303 00:23:28,499 --> 00:23:32,560 business has to see that this is not so much a problem perhaps but that it could 304 00:23:32,560 --> 00:23:36,110 be an opportunity. I think I'm still looking for a metaphor here but perhaps, 305 00:23:36,110 --> 00:23:40,350 if we you know again, compare this issue to global warming we say that we need 306 00:23:40,350 --> 00:23:44,930 something like ecological food for data. And but I don't know what that's gonna 307 00:23:44,930 --> 00:23:48,270 look like or how we're gonna explain that maybe we have to talk about fast food 308 00:23:48,270 --> 00:23:54,040 versus fast data versus ecological data but we need a metaphor here. Of course 309 00:23:54,040 --> 00:24:07,750 laws are also really helpful. So we might get things like this. I'm actually working 310 00:24:07,750 --> 00:24:18,600 on this is funny. Or if things get really out of hand we might get here, right? 311 00:24:18,600 --> 00:24:23,020 So luckily we see that in Europe the the politicians are awake and are 312 00:24:23,020 --> 00:24:25,630 really trying to push this market I think that's really great, so I think in the 313 00:24:25,630 --> 00:24:28,679 future we'll get to a moment where people say well I prefer European smart products 314 00:24:28,679 --> 00:24:33,320 for example, I think that's a good thing I think this is really positive. Finally I 315 00:24:33,320 --> 00:24:36,930 want to get to all of us what each of us can do. I think here again there's a 316 00:24:36,930 --> 00:24:40,412 parallel to global warming where at its core it's not so much about the new 317 00:24:40,412 --> 00:24:43,770 technology and all the issues, it's about a new mindset, a new way of looking at the 318 00:24:43,770 --> 00:24:48,140 world. And I here think we have to stop saying that we have nothing to hide for 319 00:24:48,140 --> 00:24:52,460 example. If I've learned anything in the past years understanding and researching 320 00:24:52,460 --> 00:24:58,200 privacy and this big trade data market is privacy is the right to be imperfect. All 321 00:24:58,200 --> 00:25:01,390 right increasing there's pressure to be the perfect citizen to be the perfect 322 00:25:01,390 --> 00:25:05,840 consumer and privacy is a way of getting out of that. So this is how I would 323 00:25:05,840 --> 00:25:09,470 reframe privacy it's not just being about which bits and bytes go where but it's 324 00:25:09,470 --> 00:25:13,040 about you know the human right to be imperfect cause course we are human we are 325 00:25:13,040 --> 00:25:17,440 all imperfect. Sometimes when I talk at technology conference people say well 326 00:25:17,440 --> 00:25:21,500 privacy was just a phase. You know, it's like ebb and flood in and we got it and 327 00:25:21,500 --> 00:25:25,980 it's gonna go away again, that's crazy you know, you don't say women's rights were 328 00:25:25,980 --> 00:25:30,710 just a phase we had it for a while and it's gonna go again. Right? And of course 329 00:25:30,710 --> 00:25:34,341 Edward Snowden explains it way better. He says arguing that you don't care about the 330 00:25:34,341 --> 00:25:37,200 right to privacy because you have nothing to hide it's no different than saying you 331 00:25:37,200 --> 00:25:40,930 don't care about free speech because you have nothing to say. What an eloquent 332 00:25:40,930 --> 00:25:47,210 system admin. So I think what we have to do strive for here is that we develop for 333 00:25:47,210 --> 00:25:50,799 more nuanced understanding of all these issues. I think we have to go away from 334 00:25:50,799 --> 00:25:54,330 this idea that data more data is better, data is automatically progress. No it's 335 00:25:54,330 --> 00:25:58,940 not data is a trade-off for example for the individual more data might mean less 336 00:25:58,940 --> 00:26:04,381 psychological security, less willingness to share, less willing to try things. For 337 00:26:04,381 --> 00:26:09,110 a country it might mean less autonomy for citizens and citizens need their own 338 00:26:09,110 --> 00:26:12,120 autonomy they need to know what's going on they need to be able to vote in their own 339 00:26:12,120 --> 00:26:17,050 autonomous way and decide what's what they want. In business you could say more data 340 00:26:17,050 --> 00:26:21,140 might lead to less creativity right less willingness to share new ideas to come up 341 00:26:21,140 --> 00:26:29,510 with new ideas - that's again an issue there. So in conclusion social cooling is 342 00:26:29,510 --> 00:26:32,460 a way of understanding these issues or a way of framing these issues that I think 343 00:26:32,460 --> 00:26:37,030 could be useful for us. That could help us understand and engage with these issues. 344 00:26:37,030 --> 00:26:41,620 And yes social cooling is an alarm, it's alarmist - it is we're trying to say this 345 00:26:41,620 --> 00:26:45,940 is the problem and we have to deal with this. But it's also really about hope. All 346 00:26:45,940 --> 00:26:50,020 right. I trust not so much in technology I trust in us in people that we can fix this 347 00:26:50,020 --> 00:26:53,530 once we understand the issue in the same way that when we understood the problem 348 00:26:53,530 --> 00:26:56,460 with global warming we started to deal with it. Where do it's 349 00:26:56,460 --> 00:27:00,030 gonna it's slow progress we're doing that and we can do the same thing with data. 350 00:27:00,030 --> 00:27:06,360 It'll take a while but we'll get there. And finally this is about starting to 351 00:27:06,360 --> 00:27:10,350 understand the difference between shallow optimism and deep optimism. All right, 352 00:27:10,350 --> 00:27:13,290 oftentimes technology sectors right cool into technology and we're going to fix 353 00:27:13,290 --> 00:27:16,570 this by creating an app and for me that's you know ,They: "we have to be 354 00:27:16,570 --> 00:27:20,880 optimistic", that's very shallow optimism the TEDx make optimism. Like true optimism 355 00:27:20,880 --> 00:27:24,480 recognizes that each technology comes with a downside and we have to recognize that 356 00:27:24,480 --> 00:27:28,420 thats it's, that thats not a problem to, to point out these problems it's a good 357 00:27:28,420 --> 00:27:32,049 thing if once you understand the problems you can deal with them - and you know come 358 00:27:32,049 --> 00:27:37,150 up with better solutions. If we don't change in this mindset then we might 359 00:27:37,150 --> 00:27:41,370 create the world where we're all more well behaved but perhaps also a little bit 360 00:27:41,370 --> 00:27:59,750 less human. Thank you. *Applause* 361 00:27:59,750 --> 00:28:04,440 H: Thank You Devin. TS: You are welcome. 362 00:28:04,440 --> 00:28:13,080 *Applause* H: We still have five more minutes we'll 363 00:28:13,080 --> 00:28:18,179 take some questions if you like. First microphone number 2. 364 00:28:18,179 --> 00:28:23,170 Microphone 2 (M2): Hello, thanks that was a really interesting talk. I have a 365 00:28:23,170 --> 00:28:29,790 question that I hope will work it's a bit complicated there's a project called indie 366 00:28:29,790 --> 00:28:34,230 by a foundation called a sovereign foundation do you know about it? Okay very 367 00:28:34,230 --> 00:28:38,720 great perfect so to just to quickly explain these people want to create an 368 00:28:38,720 --> 00:28:42,669 identity layer that will be self sovereign which means people can reveal what they 369 00:28:42,669 --> 00:28:47,530 want about themselves only when they want but is one unique identity on the entire 370 00:28:47,530 --> 00:28:51,750 internet so that can potentially be very liberating because you control all your 371 00:28:51,750 --> 00:28:57,130 identity and individual data. But at the same time it could be used to enable 372 00:28:57,130 --> 00:29:00,330 something like the personal scores we were showing earlier on so made me think about 373 00:29:00,330 --> 00:29:02,720 that and I wanted to know if you had an opinion on this. 374 00:29:02,720 --> 00:29:09,059 TS: Yes well um the first thing I think about is that as I try to explain you see 375 00:29:09,059 --> 00:29:11,440 a lot of initiatives have tried to be about: "Oo you have to control your own 376 00:29:11,440 --> 00:29:15,490 data". But that's really missing the point that it's no longer really about your data 377 00:29:15,490 --> 00:29:19,410 it's about this derived data and of course it can help to to manage what you share 378 00:29:19,410 --> 00:29:23,600 you know then they can't derive anything from it. But to little I see that 379 00:29:23,600 --> 00:29:29,360 awareness. Second of all this is very much for me an example of what nerds and 380 00:29:29,360 --> 00:29:32,480 technologies are really good at it's like: "oh we've got a social problem let's 381 00:29:32,480 --> 00:29:36,299 create a technology app and then we'll fix it". Well what I'm trying to explain is 382 00:29:36,299 --> 00:29:39,690 that this is such a big problem that we cannot fix this with just one group alone 383 00:29:39,690 --> 00:29:42,320 - not the politicians, not the designers, not the Nerds this is something that we 384 00:29:42,320 --> 00:29:47,799 have to really get together you know grab - fix together because this is such a 385 00:29:47,799 --> 00:29:50,930 fundamental issue right. The idea that risk is a problem that we want to manage 386 00:29:50,930 --> 00:29:55,570 risk is such so deeply ingrained in people you know such stuff based in fear is 387 00:29:55,570 --> 00:29:59,169 fundamental and it's everywhere so it's not enough for one group to try to fix 388 00:29:59,169 --> 00:30:02,450 that it's something that we have to come to grips with together. 389 00:30:02,450 --> 00:30:06,390 M2: Thanks a lot. H: Ok there is a signal angel has a 390 00:30:06,390 --> 00:30:09,730 question from the internet I think. Signal Angel (SigA): Yes and BarkingSheep 391 00:30:09,730 --> 00:30:13,160 is asking: "do you think there's a relationship between self-censorship and 392 00:30:13,160 --> 00:30:16,980 echo chambers in a sense that people become afraid to challenge their own 393 00:30:16,980 --> 00:30:22,799 belief and thus isolate themselves in groups with the same ideology?". 394 00:30:22,799 --> 00:30:27,830 TS: That's, a that's a, that's a really big answer to that one. 395 00:30:27,830 --> 00:30:31,440 *pauses* TS: Actually, I was e-mailing Vince Cerf, 396 00:30:31,440 --> 00:30:35,610 and miraculously he, he responded, and he said what you really have to look for is 397 00:30:35,610 --> 00:30:39,480 this, not just a reputation economy, but also the attention economy and how they're 398 00:30:39,480 --> 00:30:45,280 linked. So for a while I've been looking for that, that link and there's a lot to 399 00:30:45,280 --> 00:30:49,990 say there and there definitely is a link. I think important to understand over to 400 00:30:49,990 --> 00:30:53,540 get new ones here is that, I'm not saying that everybody will become really well 401 00:30:53,540 --> 00:30:59,700 behaved and gray book worm people. The thing is that what this situation's 402 00:30:59,700 --> 00:31:03,230 creating, is that we're all becoming theater players while playing in identity 403 00:31:03,230 --> 00:31:06,230 more and more, because we're watched more of the time. And for some people that 404 00:31:06,230 --> 00:31:10,660 might mean that they're, you know, I think most people will be more conservative and 405 00:31:10,660 --> 00:31:15,400 more careful, some people will go really all out and they all enjoy the stage! You 406 00:31:15,400 --> 00:31:18,850 know? We have those people as well, and I think those people could really benefit 407 00:31:18,850 --> 00:31:23,800 and that the attention economy could really you know give them a lot of 408 00:31:23,800 --> 00:31:26,859 attention through that. So I think there's, there's a link there but I could 409 00:31:26,859 --> 00:31:29,549 go on more but I think it's for now, where I'm aware. 410 00:31:29,549 --> 00:31:33,819 H: Okay, we're short on time, we'll take, I'm sorry one more question. The number 411 00:31:33,819 --> 00:31:36,850 one? Microphone 1 (M1): So, the, I think the 412 00:31:36,850 --> 00:31:38,710 audience you're talking about, ... H: Louder, please. 413 00:31:38,710 --> 00:31:44,350 M1: The the audience you're talking to here, is already very aware but I'm asking 414 00:31:44,350 --> 00:31:50,210 for, like tactics, or your tips, to spread your message and to talk to people that 415 00:31:50,210 --> 00:31:55,440 are in this, they say: "Uh, I don't care they can surveil me.", like what's, what's 416 00:31:55,440 --> 00:32:01,430 your approach, like in a practical way? How do you actually do this? 417 00:32:01,430 --> 00:32:08,230 TS: Yeah, so, I'm really glad to be here because I am, yes, I am a nerd, but I'm 418 00:32:08,230 --> 00:32:11,539 also a philosopher or thinker, you know and, and that means that 419 00:32:11,539 --> 00:32:15,820 for me what I work with, it's not just odd Rhinos, but words and ideas. I think those 420 00:32:15,820 --> 00:32:18,860 I've been trying to show can be really powerful, like a word can be a really 421 00:32:18,860 --> 00:32:29,229 powerful way to frame a debate or engage people. So, I haven't found yet a way to 422 00:32:29,229 --> 00:32:33,030 push all this tar. Like, I was making joke that I can tell you in one sentence, what 423 00:32:33,030 --> 00:32:36,500 privacy is and why it matters but I have to give a whole talk before that, all 424 00:32:36,500 --> 00:32:39,530 right? Privacy is a right to be imperfect but in order to understand that you have 425 00:32:39,530 --> 00:32:42,710 to understand the rise of the reputation economy, and how it affects your chances 426 00:32:42,710 --> 00:32:47,061 in life. The fun thing is, that, that that will happen by itself that people will 427 00:32:47,061 --> 00:32:50,700 become more aware of that, they will run into these problems. They will not get a 428 00:32:50,700 --> 00:32:55,660 job or they might get other issues, and then they will start to see the problem. 429 00:32:55,660 --> 00:32:59,150 And so my question not so much to help people understand it, but to help them 430 00:32:59,150 --> 00:33:02,980 understand it before they run into the wall, right? That's how usually society at 431 00:33:02,980 --> 00:33:06,530 the moment deals with technology problems. It's like "Oh we'll, we'll, oh ... Oh? 432 00:33:06,530 --> 00:33:10,820 it's a problem? Oh well, now we'll try to fix it." Well, I believe you can really 433 00:33:10,820 --> 00:33:15,020 see these problems come way earlier and I think the humanity's, to come around from, 434 00:33:15,020 --> 00:33:19,630 is really helpful in that, and trying to you know like, the lows are really, 435 00:33:19,630 --> 00:33:28,040 really clearly explaining what the problem is in 1995. So yeah, that I think that, I 436 00:33:28,040 --> 00:33:32,530 don't have a short way of explaining, you know, why privacy matters but I think 437 00:33:32,530 --> 00:33:38,220 it'll become easier over time as people start to really feel these pressures. 438 00:33:38,220 --> 00:33:42,530 H: Sorry, thank you very much for the question. I think we all should go out and 439 00:33:42,530 --> 00:33:49,320 spread the message. This talk is over, I'm awfully sorry. When you people leave, 440 00:33:49,320 --> 00:33:51,620 please take your bottles, and your cups, ... 441 00:33:51,620 --> 00:33:54,190 *applause* H: ... and all your junk, and thank you 442 00:33:54,190 --> 00:34:05,500 very much again Tijmen Schep! *applause* 443 00:34:05,500 --> 00:34:09,899 *music* 444 00:34:09,899 --> 00:34:27,000 subtitles created by c3subtitles.de in the year 2017. Join, and help us!