1 00:00:01,087 --> 00:00:03,580 So whenever I visit a school and talk to students, 2 00:00:03,604 --> 00:00:05,747 I always ask them the same thing: 3 00:00:06,754 --> 00:00:08,188 Why do you Google? 4 00:00:08,624 --> 00:00:12,021 Why is Google the search engine of choice for you? 5 00:00:12,855 --> 00:00:15,407 Strangely enough, I always get the same three answers. 6 00:00:15,431 --> 00:00:17,470 One, "Because it works," 7 00:00:17,494 --> 00:00:20,400 which is a great answer; that's why I Google, too. 8 00:00:20,424 --> 00:00:22,457 Two, somebody will say, 9 00:00:22,481 --> 00:00:25,121 "I really don't know of any alternatives." 10 00:00:25,708 --> 00:00:28,836 It's not an equally great answer and my reply to that is usually, 11 00:00:28,860 --> 00:00:30,781 "Try to Google the word 'search engine,' 12 00:00:30,805 --> 00:00:33,207 you may find a couple of interesting alternatives." 13 00:00:33,231 --> 00:00:35,326 And last but not least, thirdly, 14 00:00:35,350 --> 00:00:38,660 inevitably, one student will raise her or his hand and say, 15 00:00:38,684 --> 00:00:43,867 "With Google, I'm certain to always get the best, unbiased search result." 16 00:00:45,157 --> 00:00:51,663 Certain to always get the best, unbiased search result. 17 00:00:53,091 --> 00:00:55,481 Now, as a man of the humanities, 18 00:00:55,505 --> 00:00:57,686 albeit a digital humanities man, 19 00:00:57,710 --> 00:00:59,448 that just makes my skin curl, 20 00:00:59,472 --> 00:01:04,358 even if I, too, realize that that trust, that idea of the unbiased search result 21 00:01:04,382 --> 00:01:08,237 is a cornerstone in our collective love for and appreciation of Google. 22 00:01:08,658 --> 00:01:12,916 I will show you why that, philosophically, is almost an impossibility. 23 00:01:12,940 --> 00:01:16,194 But let me first elaborate, just a little bit, on a basic principle 24 00:01:16,218 --> 00:01:19,331 behind each search query that we sometimes seem to forget. 25 00:01:19,851 --> 00:01:21,931 So whenever you set out to Google something, 26 00:01:21,955 --> 00:01:25,882 start by asking yourself this: "Am I looking for an isolated fact?" 27 00:01:26,334 --> 00:01:29,495 What is the capital of France? 28 00:01:29,519 --> 00:01:31,944 What are the building blocks of a water molecule? 29 00:01:31,968 --> 00:01:34,309 Great -- Google away. 30 00:01:34,333 --> 00:01:37,453 There's not a group of scientists who are this close to proving 31 00:01:37,477 --> 00:01:39,474 that it's actually London and H30. 32 00:01:39,498 --> 00:01:41,869 You don't see a big conspiracy among those things. 33 00:01:41,893 --> 00:01:43,426 We agree, on a global scale, 34 00:01:43,450 --> 00:01:46,175 what the answers are to these isolated facts. 35 00:01:46,199 --> 00:01:51,501 But if you complicate your question just a little bit and ask something like, 36 00:01:51,525 --> 00:01:54,208 "Why is there an Israeli-Palestine conflict?" 37 00:01:54,978 --> 00:01:57,618 You're not exactly looking for a singular fact anymore, 38 00:01:57,642 --> 00:01:59,475 you're looking for knowledge, 39 00:01:59,499 --> 00:02:02,077 which is something way more complicated and delicate. 40 00:02:02,600 --> 00:02:04,149 And to get to knowledge, 41 00:02:04,173 --> 00:02:07,204 you have to bring 10 or 20 or 100 facts to the table 42 00:02:07,228 --> 00:02:10,204 and acknowledge them and say, "Yes, these are all true." 43 00:02:10,228 --> 00:02:11,902 But because of who I am, 44 00:02:11,926 --> 00:02:14,196 young or old, black or white, gay or straight, 45 00:02:14,220 --> 00:02:15,831 I will value them differently. 46 00:02:15,855 --> 00:02:17,543 And I will say, "Yes, this is true, 47 00:02:17,567 --> 00:02:19,681 but this is more important to me than that." 48 00:02:19,705 --> 00:02:21,695 And this is where it becomes interesting, 49 00:02:21,719 --> 00:02:23,865 because this is where we become human. 50 00:02:23,889 --> 00:02:26,885 This is when we start to argue, to form society. 51 00:02:26,909 --> 00:02:30,266 And to really get somewhere, we need to filter all our facts here, 52 00:02:30,290 --> 00:02:32,846 through friends and neighbors and parents and children 53 00:02:32,870 --> 00:02:34,902 and coworkers and newspapers and magazines, 54 00:02:34,926 --> 00:02:38,006 to finally be grounded in real knowledge, 55 00:02:38,030 --> 00:02:42,077 which is something that a search engine is a poor help to achieve. 56 00:02:43,284 --> 00:02:49,612 So, I promised you an example just to show you why it's so hard 57 00:02:49,636 --> 00:02:53,040 to get to the point of true, clean, objective knowledge -- 58 00:02:53,064 --> 00:02:54,532 as food for thought. 59 00:02:54,556 --> 00:02:58,449 I will conduct a couple of simple queries, search queries. 60 00:02:58,473 --> 00:03:02,513 We'll start with "Michelle Obama," 61 00:03:02,537 --> 00:03:04,341 the First Lady of the United States. 62 00:03:04,365 --> 00:03:06,094 And we'll click for pictures. 63 00:03:07,007 --> 00:03:09,279 It works really well, as you can see. 64 00:03:09,303 --> 00:03:12,331 It's a perfect search result, more or less. 65 00:03:12,355 --> 00:03:15,105 It's just her in the picture, not even the President. 66 00:03:15,664 --> 00:03:16,977 How does this work? 67 00:03:17,837 --> 00:03:19,209 Quite simple. 68 00:03:19,233 --> 00:03:22,448 Google uses a lot of smartness to achieve this, but quite simply, 69 00:03:22,472 --> 00:03:24,532 they look at two things more than anything. 70 00:03:24,556 --> 00:03:29,712 First, what does it say in the caption under the picture on each website? 71 00:03:29,736 --> 00:03:31,951 Does it say "Michelle Obama" under the picture? 72 00:03:31,975 --> 00:03:34,331 Pretty good indication it's actually her on there. 73 00:03:34,355 --> 00:03:36,741 Second, Google looks at the picture file, 74 00:03:36,765 --> 00:03:39,797 the name of the file as such uploaded to the website. 75 00:03:39,821 --> 00:03:42,490 Again, is it called "MichelleObama.jpeg"? 76 00:03:42,839 --> 00:03:45,761 Pretty good indication it's not Clint Eastwood in the picture. 77 00:03:45,785 --> 00:03:50,050 So, you've got those two and you get a search result like this -- almost. 78 00:03:50,074 --> 00:03:56,677 Now, in 2009, Michelle Obama was the victim of a racist campaign, 79 00:03:56,701 --> 00:04:00,716 where people set out to insult her through her search results. 80 00:04:01,430 --> 00:04:04,132 There was a picture distributed widely over the Internet 81 00:04:04,156 --> 00:04:06,800 where her face was distorted to look like a monkey. 82 00:04:06,824 --> 00:04:09,993 And that picture was published all over. 83 00:04:10,017 --> 00:04:13,778 And people published it very, very purposefully, 84 00:04:13,802 --> 00:04:15,773 to get it up there in the search results. 85 00:04:15,797 --> 00:04:18,752 They made sure to write "Michelle Obama" in the caption 86 00:04:18,776 --> 00:04:22,953 and they made sure to upload the picture as "MichelleObama.jpeg," or the like. 87 00:04:22,977 --> 00:04:25,344 You get why -- to manipulate the search result. 88 00:04:25,368 --> 00:04:26,663 And it worked, too. 89 00:04:26,687 --> 00:04:29,407 So when you picture-Googled for "Michelle Obama" in 2009, 90 00:04:29,431 --> 00:04:32,818 that distorted monkey picture showed up among the first results. 91 00:04:32,842 --> 00:04:36,408 Now, the results are self-cleansing, 92 00:04:36,432 --> 00:04:38,185 and that's sort of the beauty of it, 93 00:04:38,209 --> 00:04:41,612 because Google measures relevance every hour, every day. 94 00:04:41,636 --> 00:04:44,350 However, Google didn't settle for that this time, 95 00:04:44,374 --> 00:04:47,498 they just thought, "That's racist and it's a bad search result 96 00:04:47,522 --> 00:04:50,657 and we're going to go back and clean that up manually. 97 00:04:50,681 --> 00:04:53,613 We are going to write some code and fix it," 98 00:04:53,637 --> 00:04:54,884 which they did. 99 00:04:55,454 --> 00:04:59,196 And I don't think anyone in this room thinks that was a bad idea. 100 00:04:59,789 --> 00:05:00,953 Me neither. 101 00:05:02,802 --> 00:05:05,834 But then, a couple of years go by, 102 00:05:05,858 --> 00:05:08,842 and the world's most-Googled Anders, 103 00:05:08,866 --> 00:05:11,145 Anders Behring Breivik, 104 00:05:11,169 --> 00:05:12,875 did what he did. 105 00:05:12,899 --> 00:05:14,900 This is July 22 in 2011, 106 00:05:14,924 --> 00:05:17,573 and a terrible day in Norwegian history. 107 00:05:17,597 --> 00:05:21,384 This man, a terrorist, blew up a couple of government buildings 108 00:05:21,408 --> 00:05:24,291 walking distance from where we are right now in Oslo, Norway 109 00:05:24,315 --> 00:05:26,366 and then he traveled to the island of Utøya 110 00:05:26,390 --> 00:05:28,613 and shot and killed a group of kids. 111 00:05:29,113 --> 00:05:30,841 Almost 80 people died that day. 112 00:05:32,397 --> 00:05:36,956 And a lot of people would describe this act of terror as two steps, 113 00:05:36,980 --> 00:05:40,391 that he did two things: he blew up the buildings and he shot those kids. 114 00:05:40,415 --> 00:05:41,580 It's not true. 115 00:05:42,326 --> 00:05:44,469 It was three steps. 116 00:05:44,493 --> 00:05:46,707 He blew up those buildings, he shot those kids, 117 00:05:46,731 --> 00:05:50,375 and he sat down and waited for the world to Google him. 118 00:05:51,227 --> 00:05:53,854 And he prepared all three steps equally well. 119 00:05:54,544 --> 00:05:57,334 And if there was somebody who immediately understood this, 120 00:05:57,358 --> 00:05:58,882 it was a Swedish web developer, 121 00:05:58,906 --> 00:06:02,529 a search engine optimization expert in Stockholm, named Nikke Lindqvist. 122 00:06:02,553 --> 00:06:04,141 He's also a very political guy 123 00:06:04,165 --> 00:06:07,441 and he was right out there in social media, on his blog and Facebook. 124 00:06:07,465 --> 00:06:08,671 And he told everybody, 125 00:06:08,695 --> 00:06:11,150 "If there's something that this guy wants right now, 126 00:06:11,174 --> 00:06:13,633 it's to control the image of himself. 127 00:06:14,760 --> 00:06:16,720 Let's see if we can distort that. 128 00:06:17,490 --> 00:06:21,467 Let's see if we, in the civilized world, can protest against what he did 129 00:06:21,491 --> 00:06:24,808 through insulting him in his search results." 130 00:06:24,832 --> 00:06:26,019 And how? 131 00:06:26,797 --> 00:06:28,853 He told all of his readers the following, 132 00:06:28,877 --> 00:06:30,741 "Go out there on the Internet, 133 00:06:30,765 --> 00:06:33,660 find pictures of dog poop on sidewalks -- 134 00:06:34,708 --> 00:06:36,882 find pictures of dog poop on sidewalks -- 135 00:06:36,906 --> 00:06:40,376 publish them in your feeds, on your websites, on your blogs. 136 00:06:40,400 --> 00:06:43,321 Make sure to write the terrorist's name in the caption, 137 00:06:43,345 --> 00:06:47,832 make sure to name the picture file "Breivik.jpeg." 138 00:06:47,856 --> 00:06:51,657 Let's teach Google that that's the face of the terrorist." 139 00:06:53,552 --> 00:06:54,830 And it worked. 140 00:06:55,853 --> 00:06:58,751 Two years after that campaign against Michelle Obama, 141 00:06:58,775 --> 00:07:02,041 this manipulation campaign against Anders Behring Breivik worked. 142 00:07:02,065 --> 00:07:06,527 If you picture-Googled for him weeks after the July 22 events from Sweden, 143 00:07:06,551 --> 00:07:10,878 you'd see that picture of dog poop high up in the search results, 144 00:07:10,902 --> 00:07:12,346 as a little protest. 145 00:07:13,425 --> 00:07:17,557 Strangely enough, Google didn't intervene this time. 146 00:07:18,494 --> 00:07:22,766 They did not step in and manually clean those search results up. 147 00:07:23,964 --> 00:07:25,680 So the million-dollar question, 148 00:07:25,704 --> 00:07:29,072 is there anything different between these two happenings here? 149 00:07:29,096 --> 00:07:32,289 Is there anything different between what happened to Michelle Obama 150 00:07:32,313 --> 00:07:34,378 and what happened to Anders Behring Breivik? 151 00:07:34,402 --> 00:07:35,686 Of course not. 152 00:07:36,861 --> 00:07:38,332 It's the exact same thing, 153 00:07:38,356 --> 00:07:41,220 yet Google intervened in one case and not in the other. 154 00:07:41,244 --> 00:07:42,497 Why? 155 00:07:43,283 --> 00:07:46,583 Because Michelle Obama is an honorable person, that's why, 156 00:07:46,607 --> 00:07:49,523 and Anders Behring Breivik is a despicable person. 157 00:07:50,142 --> 00:07:51,677 See what happens there? 158 00:07:51,701 --> 00:07:54,956 An evaluation of a person takes place 159 00:07:54,980 --> 00:07:58,766 and there's only one power-player in the world 160 00:07:58,790 --> 00:08:01,270 with the authority to say who's who. 161 00:08:01,882 --> 00:08:03,623 "We like you, we dislike you. 162 00:08:03,647 --> 00:08:05,686 We believe in you, we don't believe in you. 163 00:08:05,710 --> 00:08:08,257 You're right, you're wrong. You're true, you're false. 164 00:08:08,281 --> 00:08:10,086 You're Obama, and you're Breivik." 165 00:08:10,791 --> 00:08:12,791 That's power if I ever saw it. 166 00:08:15,206 --> 00:08:18,858 So I'm asking you to remember that behind every algorithm 167 00:08:18,882 --> 00:08:20,659 is always a person, 168 00:08:20,683 --> 00:08:23,178 a person with a set of personal beliefs 169 00:08:23,202 --> 00:08:25,727 that no code can ever completely eradicate. 170 00:08:25,751 --> 00:08:28,185 And my message goes out not only to Google, 171 00:08:28,209 --> 00:08:31,019 but to all believers in the faith of code around the world. 172 00:08:31,043 --> 00:08:34,019 You need to identify your own personal bias. 173 00:08:34,043 --> 00:08:36,056 You need to understand that you are human 174 00:08:36,080 --> 00:08:38,571 and take responsibility accordingly. 175 00:08:39,891 --> 00:08:42,829 And I say this because I believe we've reached a point in time 176 00:08:42,853 --> 00:08:44,408 when it's absolutely imperative 177 00:08:44,432 --> 00:08:47,649 that we tie those bonds together again, tighter: 178 00:08:47,673 --> 00:08:50,041 the humanities and the technology. 179 00:08:50,483 --> 00:08:52,288 Tighter than ever. 180 00:08:52,312 --> 00:08:55,651 And, if nothing else, to remind us that that wonderfully seductive idea 181 00:08:55,675 --> 00:08:58,343 of the unbiased, clean search result 182 00:08:58,367 --> 00:09:01,134 is, and is likely to remain, a myth. 183 00:09:01,984 --> 00:09:03,143 Thank you for your time. 184 00:09:03,167 --> 00:09:05,599 (Applause)