WEBVTT - Frances Haugen: How One Whistle Makes a Difference 0:00:04.080 --> 0:00:07.440 Hi everyone, I'm Katie Kurk, and this is next question. 0:00:08.119 --> 0:00:10.640 In September of twenty twenty one, The Wall Street Journal 0:00:10.680 --> 0:00:13.280 began to roll out a series of eleven stories that 0:00:13.320 --> 0:00:16.200 would have a major impact on the way people think 0:00:16.239 --> 0:00:20.400 about technology companies. The series was called The Facebook Files, 0:00:20.600 --> 0:00:23.200 and the source for the series maybe one of the 0:00:23.239 --> 0:00:27.640 biggest journalistic treasure troves of the century, was a former 0:00:27.720 --> 0:00:32.600 Facebook product manager named Francis Hogan. Francis had only been 0:00:32.640 --> 0:00:35.040 at Facebook for a couple of years, but during that 0:00:35.159 --> 0:00:39.440 time she became increasingly alarmed by a disturbing pattern. 0:00:39.800 --> 0:00:41.920 The only data that we get out of these companies 0:00:42.000 --> 0:00:44.040 is how many users do they have? How much time 0:00:44.040 --> 0:00:45.680 do you spend, how the ads do you book on? 0:00:45.760 --> 0:00:49.600 What's that revenue? You don't get the societal costs that 0:00:49.680 --> 0:00:51.159 come as a consequence. 0:00:51.600 --> 0:00:55.400 It seemed that Facebook was prioritizing their own profits over 0:00:55.480 --> 0:01:00.720 public safety and putting people's lives at risk. So she 0:01:00.840 --> 0:01:04.320 blew the whistle. She made tens of thousands of pages 0:01:04.360 --> 0:01:07.880 of internal documents available to The Wall Street Journal. And 0:01:08.000 --> 0:01:11.760 what happened next testifying before the US Congress, the UK 0:01:11.959 --> 0:01:16.000 and EU parliaments, and filing a complaint with the sec 0:01:16.760 --> 0:01:20.000 exposed the depths that the tech company would go to 0:01:20.000 --> 0:01:23.960 to mislead the public and grow its bottom line. Now, 0:01:24.000 --> 0:01:27.119 Francis has written a book. It's called The Power of One. 0:01:27.480 --> 0:01:30.040 How I found the strength to tell the truth and 0:01:30.160 --> 0:01:33.560 why I blew the whistle on Facebook, And I'm very 0:01:33.600 --> 0:01:37.520 glad that it's brought her to next question. First of all, Francis, 0:01:37.760 --> 0:01:41.000 it's great to finally meet you. I am a huge 0:01:41.040 --> 0:01:42.920 admirer of yours, so. 0:01:42.800 --> 0:01:44.880 Excited to be here. Thank you so much for inviting me. 0:01:45.280 --> 0:01:48.639 Of course, it's been quite a ride for you since 0:01:48.800 --> 0:01:54.440 you became the Facebook whistleblower in October of twenty twenty one, 0:01:54.560 --> 0:01:57.160 now a year and a half or so later. How 0:01:57.160 --> 0:02:00.760 are you feeling? Are you happy that you came forward? 0:02:00.880 --> 0:02:02.880 Do you have any regrets? Would you do it all 0:02:02.920 --> 0:02:03.559 over again? 0:02:04.240 --> 0:02:07.120 You know, when I originally came forward, I had very 0:02:07.240 --> 0:02:10.200 very basic goals, like I wanted to not have to 0:02:10.240 --> 0:02:12.760 carry a secret that I thought had the potential to 0:02:12.840 --> 0:02:15.960 really impact the lives of others. I came forward because 0:02:16.080 --> 0:02:19.160 I was concerned about how Facebook was operating in African 0:02:19.160 --> 0:02:23.320 countries in Southeast Asia, and I then and still genuinely 0:02:23.360 --> 0:02:25.880 believe that if we continue to operate the way we do, 0:02:26.280 --> 0:02:28.440 there are millions of lives on the line from things 0:02:28.480 --> 0:02:31.760 like ethnic violence. But the world has changed a lot 0:02:31.840 --> 0:02:34.240 since I came out. I was really shocked last week 0:02:34.280 --> 0:02:37.440 when the Surgeon General issued his advisory on social media 0:02:37.480 --> 0:02:40.840 and mental health for kids. I've been amazed at how 0:02:40.919 --> 0:02:46.000 just knowing that these companies knew these harms were real 0:02:46.080 --> 0:02:49.520 across a wide variety of harms has really galvanized the 0:02:49.560 --> 0:02:53.640 activist community. It's caused legislative conversations around the world that 0:02:54.000 --> 0:02:57.120 were pretty stalled for a long time. And so if 0:02:57.160 --> 0:02:58.919 I could do it again, I would totally do it again. 0:02:59.160 --> 0:03:02.560 I've been incredibly fortunate and how smooth it's gone, and 0:03:02.760 --> 0:03:04.840 it's exceeded my wildest expectations. 0:03:05.160 --> 0:03:08.040 You've written a book about your experiences. Why did you 0:03:08.160 --> 0:03:13.200 want to take pen to paper, fingers to laptop and 0:03:13.400 --> 0:03:15.000 share your story with the world. 0:03:16.280 --> 0:03:18.800 One of the things that kind of baffles tech journalists 0:03:18.800 --> 0:03:22.160 when I talk to them is because tech journalists live 0:03:22.200 --> 0:03:23.920 in a little bit of an echo chamber. You know, 0:03:24.000 --> 0:03:26.520 like our classic criticism of tech is tech lives in 0:03:26.520 --> 0:03:30.480 an echo chamber. Tech journalists also live in an echo chamber. 0:03:30.520 --> 0:03:32.080 A little bit when I say to them, you know, 0:03:32.320 --> 0:03:35.120 when I take flights. You know, I'm a friendly person. 0:03:35.360 --> 0:03:37.400 I'm one of those annoying people that talks to their 0:03:37.400 --> 0:03:38.000 seed mates. 0:03:38.280 --> 0:03:39.000 I did that too. 0:03:39.200 --> 0:03:43.000 Yeah, and it's amazing. At least half the people I 0:03:43.000 --> 0:03:45.920 sit next to you have never heard of the Facebook whistleblower. 0:03:46.400 --> 0:03:48.840 And so it's one of these things where culture change, 0:03:48.960 --> 0:03:51.360 and that's the thing that we really need. We need 0:03:51.400 --> 0:03:54.640 to reset our relationship with these companies. It takes a 0:03:54.760 --> 0:03:59.480 long time, and this book, I'm hoping helps a much 0:03:59.560 --> 0:04:02.080 much larger set of people, much more diverse set of 0:04:02.120 --> 0:04:04.800 people get a seat at the table, but kind of 0:04:04.880 --> 0:04:08.120 laying out, like what are the conversations we need to 0:04:08.120 --> 0:04:10.480 be having, Like what are the choices we get to 0:04:10.520 --> 0:04:12.600 make in the next few years, because we are in 0:04:12.640 --> 0:04:15.000 a moment of inflection and we need to have as 0:04:15.040 --> 0:04:17.000 many informed people at the table as possible. 0:04:17.320 --> 0:04:19.960 You know, you talk about the Vic Murphy, the Surgeon 0:04:20.080 --> 0:04:23.920 General issuing a warning about the dangers of social media 0:04:24.240 --> 0:04:29.000 for kids. Other people becoming more aware, and I'm just 0:04:29.120 --> 0:04:32.240 curious if you feel this way. But for me, I'm 0:04:32.320 --> 0:04:35.719 kind of like, what took so long? This doesn't take 0:04:35.760 --> 0:04:39.039 a brain surgeon, a rocket scientist, or a tech expert 0:04:39.480 --> 0:04:43.680 to know that people were becoming have become incredibly addicted 0:04:44.080 --> 0:04:48.000 to social media. Tristan Harris was sounding the alarm after 0:04:48.040 --> 0:04:51.880 he left Google. All kinds of experts were saying, this 0:04:52.160 --> 0:04:56.000 is dangerous. Why did it take so long for this 0:04:56.200 --> 0:04:59.200 to really become headline news. 0:05:00.200 --> 0:05:02.320 One of the things I talk about in my book is, 0:05:02.400 --> 0:05:05.240 you know, what's the difference between say, the car industry 0:05:05.240 --> 0:05:08.200 and like the automotive industry and social media when it 0:05:08.240 --> 0:05:12.320 comes to our ability to hold it accountable or or 0:05:12.360 --> 0:05:16.240 our ability to understand it. Back in nineteen sixty five, 0:05:16.480 --> 0:05:19.760 it's going to sound shocking. There were no seatbelts in cars, 0:05:20.160 --> 0:05:23.359 no airbags. I remember that, Yeah, Like I listened to 0:05:23.400 --> 0:05:25.719 like my parents still stories about the kids all like 0:05:25.880 --> 0:05:27.600 jumbling all over each other in the back of the 0:05:27.920 --> 0:05:31.080 station wagon, and it's like, really, wow, a different world. 0:05:31.760 --> 0:05:33.680 We now put eight year olds in car seats, right. 0:05:34.040 --> 0:05:36.599 But the world changed very suddenly when a guy named 0:05:36.680 --> 0:05:39.400 Ralph Nader came out with a book called Unsafe at 0:05:39.440 --> 0:05:43.360 Any Speed. And what really changed was that people didn't 0:05:43.400 --> 0:05:46.400 realize that there was the ability to live in a 0:05:46.440 --> 0:05:49.720 different world. You know, that our fatality rate today is 0:05:50.120 --> 0:05:53.040 way less per mild driven for cars because of a 0:05:53.160 --> 0:05:55.760 long series of actions. But the thing that people need 0:05:55.800 --> 0:05:58.479 to understand is when when Ralph Nator published that book. 0:05:58.839 --> 0:06:01.800 You know, there were one hundred thousand automotive engineers in 0:06:01.839 --> 0:06:05.000 the world when I came forward. I think they were 0:06:05.000 --> 0:06:07.320 on in order, probably three hundred or four hundred people 0:06:07.360 --> 0:06:10.760 in the world who really understood how systems like facebooks work. 0:06:11.640 --> 0:06:15.160 And of those people, you know, we are educated in 0:06:15.200 --> 0:06:17.960 such narrow ways, I think a lot of those people 0:06:18.040 --> 0:06:22.440 didn't understand the larger societal consequences of those choices and decisions. 0:06:22.920 --> 0:06:26.640 And so Ralph Nader could have a chorus of automotive 0:06:26.680 --> 0:06:30.040 engineers all say this is happening. When it comes to 0:06:30.080 --> 0:06:33.400 social media, each of us sees a different world. You know, 0:06:33.520 --> 0:06:36.599 for many, many, many people who would be the ones 0:06:36.640 --> 0:06:39.920 asking those questions. When they open social media, they see 0:06:39.960 --> 0:06:43.880 their friends and family who are likely relatively similar to themselves. 0:06:44.360 --> 0:06:47.320 You know, the idea that Facebook could be radically different, 0:06:47.480 --> 0:06:50.719 radically more dangerous in a place like an African country 0:06:51.120 --> 0:06:53.520 or in Southeast Asia, it sounds foreign to us. Were 0:06:53.560 --> 0:06:55.880 like social media is about looking at pictures of cats, 0:06:56.440 --> 0:06:58.159 and so I think that's a big part of it, 0:06:58.240 --> 0:07:00.440 Like we need to be able to have the right 0:07:00.520 --> 0:07:03.240 to study social media, We need to have the right 0:07:03.279 --> 0:07:05.880 to be able to get independent data off these systems, 0:07:06.160 --> 0:07:08.400 because then we can have definitive conversations. 0:07:08.760 --> 0:07:12.600 Is that because we're not having a universal experience. It's 0:07:12.640 --> 0:07:18.840 a highly deeply personalized experience for everyone. So it's not 0:07:18.960 --> 0:07:23.800 as if we're all driving cars. We're all on different vehicles, 0:07:23.840 --> 0:07:28.280 if you will. So it's not unifying people to realize 0:07:28.280 --> 0:07:30.040 that they have to demand change. 0:07:31.400 --> 0:07:35.000 It's really important for people to understand just how different 0:07:35.480 --> 0:07:38.200 those worlds are in terms of transparency. And that's part 0:07:38.200 --> 0:07:39.960 of why I wrote the book. You know, I've I've 0:07:39.960 --> 0:07:42.040 got to live a lot of that arc of how 0:07:42.040 --> 0:07:44.240 we write software or like what does it mean to 0:07:44.240 --> 0:07:47.680 have experiences online? And I wanted to walk people through, 0:07:48.160 --> 0:07:50.239 you know, this is what changed from step to step 0:07:50.280 --> 0:07:52.760 to step so that more people have that context. 0:07:52.880 --> 0:07:56.239 I'm just curious, Francis, I know you weren't super psyched 0:07:56.240 --> 0:08:00.240 to go to Facebook. When did you realize we're not 0:08:00.280 --> 0:08:03.440 in Kansas anymore? Something is awry. 0:08:03.600 --> 0:08:05.400 It was interesting, like when they reached out to me, 0:08:05.440 --> 0:08:07.640 I was like the only thing I work on is misinformation. 0:08:08.160 --> 0:08:10.440 You know. It's interesting. I got there and I think 0:08:10.520 --> 0:08:13.080 one of the first moments where I was like, wow, 0:08:13.200 --> 0:08:16.600 this is this is chaotic is to the role I 0:08:16.640 --> 0:08:19.920 had is something called a product manager. So product managers 0:08:20.440 --> 0:08:24.520 are responsible for helping articulate what is the problem we're 0:08:24.520 --> 0:08:27.960 trying to solve, how might we solve that, and then 0:08:28.160 --> 0:08:30.880 once we come to consensus on a solution, what's this 0:08:31.080 --> 0:08:34.280 series of engineering tasks that will allow us to execute 0:08:34.280 --> 0:08:37.280 that solution. I had a role of being a product manager. 0:08:37.559 --> 0:08:41.880 And Facebook understood that they were a different enough company 0:08:42.800 --> 0:08:44.679 that they had seen that if people came in from 0:08:44.720 --> 0:08:48.240 the outside, they didn't succeed at a very high rate, 0:08:48.280 --> 0:08:50.280 like there was a lot of churn, and so they 0:08:50.480 --> 0:08:53.600 established a boot camp for two full weeks to just 0:08:53.720 --> 0:08:56.080 give kind of a basic level of like heire's how 0:08:56.120 --> 0:08:59.560 Facebook works. And my manager pulled me out of it 0:08:59.600 --> 0:09:02.040 with a like three days. He was like, we're you know, 0:09:02.080 --> 0:09:04.199 things aren't fired too much, Like we have to come 0:09:04.280 --> 0:09:06.120 up with a plan for the next six months, even 0:09:06.200 --> 0:09:08.640 though you know nothing about the problem or what's going on, 0:09:08.880 --> 0:09:11.439 Like we need you an articulative plan. Now. That was 0:09:11.520 --> 0:09:14.079 kind of my first warning that I was like, oh, wow, 0:09:14.200 --> 0:09:16.840 like the house is on fire. The house is on fire, 0:09:16.840 --> 0:09:19.559 and people are running around, even having the self awareness 0:09:19.640 --> 0:09:22.760 to be like, oh, we know that if people don't 0:09:22.760 --> 0:09:25.600 get at least a certain amount of bootstrapping. Facebook is 0:09:25.760 --> 0:09:28.120 very hard to figure out, even internally, how it works. 0:09:28.880 --> 0:09:31.120 And it was interesting. I showed up for that first meeting, 0:09:31.160 --> 0:09:32.960 you know, the one that my managers like urgently pushing 0:09:32.960 --> 0:09:36.840 me to prepare for, and we spent twenty minutes basically 0:09:36.840 --> 0:09:40.079 discussing should I have a job? So imagine you show up, 0:09:40.480 --> 0:09:43.079 You've just gotten hired for this thing, and all the 0:09:43.160 --> 0:09:45.360 leadership you're saying is like, why why do you have 0:09:45.400 --> 0:09:47.920 a team? Like why does this team exist? Like think 0:09:47.920 --> 0:09:51.200 about that for a moment. You know, because activists have 0:09:51.280 --> 0:09:54.320 told you, because the un told you, Hey, and me 0:09:54.400 --> 0:09:58.800 and Mark, your negligence around misinformation killed twenty four thousand people. 0:09:59.559 --> 0:10:02.000 You know of a problem. And yet I could sit 0:10:02.040 --> 0:10:04.680 in a room full of the leadership of safety having 0:10:04.720 --> 0:10:08.079 them be like shouldn't this team exist? Right? You can 0:10:08.120 --> 0:10:10.079 imagine that first six months was a little stressful. 0:10:10.880 --> 0:10:13.880 What was Facebook doing in the lead up to the events? 0:10:13.920 --> 0:10:17.720 On January sixth, we'll talk about that right after this, 0:10:22.440 --> 0:10:26.480 we're back with Francis Hogan, Facebook whistleblower and author of 0:10:26.520 --> 0:10:33.360 the Power of Want. Francis, did Facebook really care about 0:10:33.400 --> 0:10:37.480 misinformation or did the company just feel like dealing with 0:10:37.559 --> 0:10:39.600 it was an exercise and futility. 0:10:39.840 --> 0:10:42.319 I think part of the problem was that Facebook had 0:10:42.320 --> 0:10:45.920 taken the most obvious path to deal with misinformation, which was, 0:10:46.400 --> 0:10:49.840 let's hire experts, let's hire journalists to help us assess 0:10:49.880 --> 0:10:53.160 what's true and false. But that kind of approach, you know, 0:10:53.280 --> 0:10:56.840 fixing safety after the fact, like, oh, we've already hyper 0:10:56.840 --> 0:10:59.720 amplified you know, extreme content. Now we're going to pluck 0:10:59.720 --> 0:11:03.360 out the dangerous parts. Those strategies don't scale. You know, 0:11:03.400 --> 0:11:07.160 Facebook has three billion users. There were maybe a few 0:11:07.200 --> 0:11:10.600 thousand fact checks being done a month, maybe a few thousand. 0:11:11.000 --> 0:11:12.800 You can see why you actually need to take a 0:11:12.800 --> 0:11:15.120 different kind of approach, which is coming in and saying, 0:11:15.880 --> 0:11:19.720 why are the algorithms rewarding extreme content? How do we 0:11:19.840 --> 0:11:22.360 change our operations to deal with that fact? 0:11:23.000 --> 0:11:26.240 As they say, lies make it around the world before 0:11:26.360 --> 0:11:29.280 the truth has the chance to tie its shoes, And 0:11:29.400 --> 0:11:33.000 so what was the alternative? What would that approach it? 0:11:33.080 --> 0:11:35.120 Ben I mean, what is the answer. 0:11:35.679 --> 0:11:38.199 Let's take, for example, something as simple as should you 0:11:38.280 --> 0:11:40.640 have to click on a link before you reshare it, 0:11:41.280 --> 0:11:43.360 so you hit the nail on the head. One of 0:11:43.400 --> 0:11:46.800 the problems with third party fact checking is journalism takes time. 0:11:47.280 --> 0:11:49.880 At Facebook, on average, it was like two or three 0:11:50.000 --> 0:11:52.199 days for someone to write a fact check, and they 0:11:52.280 --> 0:11:54.240 put a huge amount of effort into trying to build 0:11:54.280 --> 0:11:58.280 prediction systems to guess which are the pieces of content 0:11:58.360 --> 0:12:01.760 that might go viral, because we have to give the 0:12:01.840 --> 0:12:04.920 journalists a head start. Alternatives are things like if you 0:12:05.000 --> 0:12:07.439 require people to click on a link before they reshare it, 0:12:07.720 --> 0:12:11.320 that reduces misinformation by like ten or fifteen percent, just 0:12:11.320 --> 0:12:13.839 because people have to pause and think for a moment. 0:12:14.480 --> 0:12:17.680 You know, there's one or two billion people on Facebook 0:12:17.720 --> 0:12:20.920 who live in places where Facebook is the internet. You know, 0:12:20.960 --> 0:12:23.880 they might have become literate to use Facebook, and as 0:12:23.920 --> 0:12:27.560 a result those places. In some countries, thirty five percent 0:12:27.640 --> 0:12:30.360 of everything you see on your news feed is a 0:12:30.440 --> 0:12:33.920 reshare and so Facebook wasn't willing to take the hit 0:12:34.120 --> 0:12:36.680 of you know, zero point one point two percent less 0:12:36.679 --> 0:12:39.880 profit of reducing the amount of content that was moving 0:12:39.880 --> 0:12:42.720 through the ecosystem. As a whole. So those are kinds 0:12:42.760 --> 0:12:45.880 of product design ways of dealing with misinformation. 0:12:46.520 --> 0:12:50.760 Our third party fact checker still the primary way Facebook 0:12:50.840 --> 0:12:54.040 is ostensibly trying to combat miss and disinformation. 0:12:54.760 --> 0:12:57.520 You know, I don't know like because we have no transparency. 0:12:58.000 --> 0:13:00.440 You know, right now, we don't have any trend parency 0:13:00.480 --> 0:13:03.640 into how these book operates. We know that Mark fired 0:13:03.720 --> 0:13:06.440 lots and lots of safety people during his year of efficiency, 0:13:06.960 --> 0:13:09.760 so it's possible that things have changed, but given his 0:13:09.960 --> 0:13:13.080 recent behavior, it's unlikely things have materially changed. 0:13:13.600 --> 0:13:19.160 Let's talk about sort of the public facing explanation from Facebook. 0:13:19.240 --> 0:13:23.120 I interviewed Cheryl Samberg in twenty nineteen when you were 0:13:23.160 --> 0:13:25.400 still working at Facebook. I asked her if she felt 0:13:25.480 --> 0:13:29.280 like Facebook was doing enough to invest in its security. 0:13:29.760 --> 0:13:34.160 Let's take a listen and let's hear your reaction to 0:13:34.160 --> 0:13:35.200 what Cheryl told me. 0:13:35.880 --> 0:13:39.760 We've put tremendous engineering resources, and we're doing things like 0:13:39.880 --> 0:13:42.120 red teams, asking what do we think the bad guys 0:13:42.160 --> 0:13:44.520 would do and how would we do it? So we're 0:13:44.559 --> 0:13:47.040 never going to fully be ahead of everything. But if 0:13:47.080 --> 0:13:48.959 you look at if you want to understand what companies 0:13:49.040 --> 0:13:51.400 care about, you look at where they invest their resources 0:13:51.640 --> 0:13:53.480 and if you look back three to five years and 0:13:53.520 --> 0:13:56.520 you look at today, we've totally changed when we invest 0:13:56.520 --> 0:13:59.559 our resources. And my job has changed too. If I 0:13:59.559 --> 0:14:01.880 look at up at Facebook eleven and a half years, 0:14:02.360 --> 0:14:05.800 for the first eight or so, I spent most of 0:14:05.840 --> 0:14:08.800 my time growing the company and sometime protecting the community. 0:14:08.800 --> 0:14:11.920 We always did some protection, but now that's definitely flipped. 0:14:11.960 --> 0:14:15.640 My job is a majority building the systems that protect 0:14:15.720 --> 0:14:19.280 and minority grow. And so we're definitely changing as a company. 0:14:19.560 --> 0:14:21.440 We're in a different place across the board on all 0:14:21.480 --> 0:14:22.080 of these things. 0:14:22.160 --> 0:14:24.200 Do you think you're changing enough fast enough? 0:14:25.360 --> 0:14:26.560 I hope, so we're trying. 0:14:27.080 --> 0:14:29.640 What's your reaction to that, Francis, It's. 0:14:29.520 --> 0:14:33.040 So interesting when we listen to media, you know, a 0:14:33.080 --> 0:14:35.600 film clip, an audio clip from the past, you know, 0:14:35.640 --> 0:14:38.400 you can often hear the emotional echoes of that moment. 0:14:38.920 --> 0:14:41.400 And I think back in twenty nineteen, she's quite earnest, 0:14:41.680 --> 0:14:44.480 like the only part of Facebook that was growing was 0:14:44.560 --> 0:14:47.320 the civic Integrity team. I think twenty nineteen was the 0:14:47.400 --> 0:14:50.280 year that the UN report on Menmark came out and 0:14:50.360 --> 0:14:53.440 you know, firmly placed blame on Facebook. They were still 0:14:53.440 --> 0:14:57.000 living right in the immediate echoes of Cambridge Analytico for context, 0:14:57.120 --> 0:14:59.400 right about when when you interviewed her, that would have 0:14:59.440 --> 0:15:03.640 been when the FTC find Facebook five billion dollars because 0:15:03.640 --> 0:15:07.360 of privacy violations from Cambridge Alica. But over the course 0:15:07.400 --> 0:15:09.440 of the next two years or even the next year, 0:15:09.920 --> 0:15:13.720 I think Facebook began to realize that having a big 0:15:13.720 --> 0:15:19.360 safety team, having people with PhDs asking questions was putting 0:15:19.400 --> 0:15:23.520 Facebook in a quite awkward position because the more people 0:15:23.880 --> 0:15:27.280 dug in, the more people having the ability to ask questions, 0:15:27.720 --> 0:15:30.960 they found things, and they found things that were quite disturbing. 0:15:31.400 --> 0:15:33.320 I think they did a pretty good job in the 0:15:33.400 --> 0:15:36.480 run up to the twenty twenty election, but as soon 0:15:36.480 --> 0:15:39.800 as the twenty twenty election passed, they fired that team, 0:15:39.960 --> 0:15:41.160 the Civic Integrity team. 0:15:41.280 --> 0:15:44.680 One month after the election, yeah, one month, and of 0:15:44.720 --> 0:15:49.200 course five weeks later was when the Trump supporters, many 0:15:49.240 --> 0:15:53.880 of whom organized on Facebook, stormed the US capital. 0:15:54.440 --> 0:15:57.160 And I think part of what happened was their papers 0:15:57.160 --> 0:16:00.240 in the Facebook files. They say like, we saw all 0:16:00.280 --> 0:16:05.000 this building up. But I think because now no single 0:16:05.080 --> 0:16:08.480 person was responsible, no one felt like they had the 0:16:08.520 --> 0:16:10.600 authority to go in there and intervene. 0:16:10.920 --> 0:16:13.320 Where do you draw the line? And nobody has a 0:16:13.320 --> 0:16:16.440 crystal ball to say these people are going to do that. 0:16:16.920 --> 0:16:20.960 Facebook talks about the difference between movements and we are 0:16:20.960 --> 0:16:25.080 to knows adversarial movements. So an adversarial movement knows that 0:16:25.120 --> 0:16:29.800 they're violating Facebook's policies and actively does countermeasures to try 0:16:29.840 --> 0:16:33.280 to get around Facebook. That's one we differentiate between, like, 0:16:33.480 --> 0:16:36.840 does this movement think they're doing something wrong? Right? 0:16:37.280 --> 0:16:37.520 Right? 0:16:37.760 --> 0:16:39.960 And you saw that extensively with Stop the Seal. 0:16:40.360 --> 0:16:44.560 If the Facebook Civic Integrity Team had not been disbanded 0:16:44.880 --> 0:16:48.320 less than one month after the election, what would they 0:16:48.360 --> 0:16:51.320 have seen and what would they have done with the 0:16:51.400 --> 0:16:54.520 activity they witnessed going on on the platform. 0:16:54.960 --> 0:16:56.720 In the round of the twenty twenty election, there were 0:16:56.720 --> 0:16:58.880 a number of things in place where Facebook said, hey, 0:16:58.960 --> 0:17:01.560 we know we have vulnerability in our system. For example, 0:17:02.200 --> 0:17:04.840 live videos. This is where you know, I can film 0:17:04.880 --> 0:17:07.680 something on my phone and Facebook will put a little 0:17:07.680 --> 0:17:11.040 announcement at the top of people's feeds. Facebook knew that 0:17:11.119 --> 0:17:14.680 live video was a particularly big vulnerability for the company 0:17:14.840 --> 0:17:19.040 because video is harder to monitor than audio or text, 0:17:19.080 --> 0:17:21.680 for sure. So you either can deal with that after 0:17:21.680 --> 0:17:24.760 the fact, or you can say, hey, what's leading to 0:17:25.119 --> 0:17:28.000 that video? Going viral, and in the case of live video, 0:17:28.400 --> 0:17:31.159 Facebook said, hey, you know, every piece of content on 0:17:31.200 --> 0:17:35.680 Facebook earns a score based on how relevant it is 0:17:35.960 --> 0:17:38.399 to uktie or to your listener. You know, is it 0:17:38.440 --> 0:17:40.520 similar to other things that they've seen before. Does this 0:17:40.560 --> 0:17:43.239 person generally produce content that people would like to engage with? 0:17:43.320 --> 0:17:45.560 You know, there's a bunch of factors. You earn a score, 0:17:45.880 --> 0:17:48.320 and that gives you a priority in the news feed. 0:17:48.800 --> 0:17:51.440 When it came to live video, they would give a boost. 0:17:51.560 --> 0:17:53.760 They'd say that score, We're going to multiply it by 0:17:53.800 --> 0:17:56.760 eight hundred and fifty times to make sure that it 0:17:56.760 --> 0:17:59.880 will show up at the top of your feed. They said, hey, 0:18:00.160 --> 0:18:02.399 we know this is dangerous. We're going to only boost 0:18:02.440 --> 0:18:06.159 it sixty five times. In the runchy election, it's a 0:18:06.240 --> 0:18:09.679 little tiny detail, but when they stormed the capitol, the 0:18:09.800 --> 0:18:13.120 rioters actively used live video to coordinate, and so there's 0:18:13.160 --> 0:18:15.560 these little things where they could have had the safety 0:18:15.600 --> 0:18:19.199 measures on that were on election day, but because no 0:18:19.240 --> 0:18:21.480 one felt they had the authority to say we're in 0:18:21.520 --> 0:18:24.720 a situation, no one turned those on until the day 0:18:24.800 --> 0:18:28.119 after they stormed the capital. These are little, tiny details, 0:18:28.480 --> 0:18:31.879 but you have to remember When people interviewed the rioters 0:18:31.920 --> 0:18:36.440 after January six, they said it seemed real. It seemed real. 0:18:36.480 --> 0:18:39.280 It seemed like everyone was saying, like, we're about to 0:18:39.359 --> 0:18:41.959 experience a coup, like we need to like go and 0:18:42.000 --> 0:18:46.159 save our democracy. These little product tweaks would have changed 0:18:46.160 --> 0:18:49.360 the information environment that those people experienced, and who knows 0:18:49.400 --> 0:18:50.960 what would have happened with January six. 0:18:51.400 --> 0:18:53.919 So why didn't they do it? Because the team had 0:18:53.960 --> 0:18:54.720 been dissolved. 0:18:55.560 --> 0:18:58.080 There was no longer a person in the company who 0:18:58.320 --> 0:19:00.359 wore the hat of saying, let's make sure we're a 0:19:00.359 --> 0:19:03.840 positive force in society, right, there was diffuse responsibility for 0:19:03.920 --> 0:19:07.600 Little Tiny Slippers. And I think after they dissolved the team, 0:19:07.600 --> 0:19:10.880 which was on December second or December third, I don't 0:19:10.880 --> 0:19:12.879 think there was anyone who felt like they had the 0:19:12.920 --> 0:19:15.720 authority to say, hey, some people are going to have 0:19:15.720 --> 0:19:18.119 to work over the holidays, right, Like this is a 0:19:18.119 --> 0:19:20.000 big enough deal that someone's going to have to do something. 0:19:20.520 --> 0:19:22.560 And I think that's why Facebook was asleep at the wheel. 0:19:23.320 --> 0:19:26.080 When we come back, Francis and I talk about improvements 0:19:26.080 --> 0:19:29.560 in social media that could have a positive impact on 0:19:29.640 --> 0:19:39.520 the team Mental health crisis. We're back with Francis, Howgan 0:19:42.160 --> 0:19:46.280 do you think Mark Zuckerberg just cares about profit over everything? 0:19:46.359 --> 0:19:50.000 And is there something about the broader culture of Facebook 0:19:50.359 --> 0:19:57.000 that makes this almost a Sisifian task to try to 0:19:57.080 --> 0:20:03.360 control or at least even monit or remove really dangerous content. 0:20:04.480 --> 0:20:07.240 So I'm glad that you bring up Mark, so just 0:20:07.280 --> 0:20:10.879 for people understand how different the leadership of Facebook is 0:20:10.960 --> 0:20:15.320 versus other companies. Mark Zuckerberg holds about fifty five fifty 0:20:15.359 --> 0:20:18.760 six percent of the voting shares that control Facebook, so 0:20:18.800 --> 0:20:21.359 that means he's the chairman of the board, He's the CEO. 0:20:22.040 --> 0:20:25.359 If he wants to invest tens of billions of dollars 0:20:25.400 --> 0:20:27.919 in the metaverse, no one can stop him because he 0:20:28.040 --> 0:20:32.000 is the only voice that matters. I do think responsibility 0:20:32.040 --> 0:20:34.359 goes to the top, right, Like part of the challenge 0:20:34.359 --> 0:20:36.720 here is you have a man who has been CEO 0:20:36.840 --> 0:20:40.800 since he was nineteen years old. Facebook is intimately tied 0:20:40.840 --> 0:20:43.359 to his identity, and it's very hard for people to 0:20:43.400 --> 0:20:46.040 admit that their life's work might be hurting other people. 0:20:46.680 --> 0:20:48.960 And so unfortunately, there is an internal culture to the 0:20:49.000 --> 0:20:53.480 company where the people who surround Mark know that being 0:20:53.520 --> 0:20:56.040 too critical isn't going to get you very far, I said, 0:20:56.040 --> 0:20:58.520 I think part of why Cheryl left from when I 0:20:58.560 --> 0:21:02.040 came out, she left maybe six once after the Facebook 0:21:02.080 --> 0:21:04.760 files happened. I think as Cheryl was a voice that 0:21:04.920 --> 0:21:08.000 was trying to push for responsibility, and there wasn't really 0:21:08.040 --> 0:21:09.840 an appetite in Tronto the company to do that. 0:21:10.280 --> 0:21:13.639 To this point, in twenty nineteen, I spoke with her 0:21:13.720 --> 0:21:18.960 about whether Facebook's business model ultimately rendered implementing security measures 0:21:19.040 --> 0:21:22.040 bad for business. Let's hear what she said. 0:21:22.640 --> 0:21:25.560 So on this, I'm really pretty proud of our track record. 0:21:25.640 --> 0:21:28.480 If you look a number of years ago and you 0:21:28.560 --> 0:21:31.159 listen to our earnings calls, So earnings calls are exactly 0:21:31.240 --> 0:21:33.600 what people are worried about. They're directed at investors. It's 0:21:33.600 --> 0:21:36.119