WEBVTT - Inside The Facebook Oversight Board 0:00:15.396 --> 0:00:33.356 Pushkin Pushkin from Pushkin Industries. This is deep background to 0:00:33.476 --> 0:00:36.796 show where we explore the stories behind the stories in 0:00:36.836 --> 0:00:41.516 the news. One of the recurring stories is Facebook, and 0:00:41.716 --> 0:00:45.476 particularly criticisms of Facebook that have arisen from a large 0:00:45.516 --> 0:00:49.236 body of material leaked by a whistleblower to the press 0:00:49.476 --> 0:00:52.836 and to Congress. This series of leaks has, of course 0:00:52.956 --> 0:00:56.556 major implications for the company, but it also has implications 0:00:56.756 --> 0:01:01.756 for the Oversight Board associated with Facebook, the independent institution 0:01:01.996 --> 0:01:05.556 that has the job of reviewing the company's most controversial 0:01:05.756 --> 0:01:09.636 content decisions and correcting them if it believes they've gotten 0:01:09.676 --> 0:01:14.076 them wrong. Regular listeners of the podcast know that I'm 0:01:14.076 --> 0:01:16.796 interested in the Oversight Board, but I would take it 0:01:16.796 --> 0:01:19.716 a bit further. I am personally interested in the Oversight Board, 0:01:19.876 --> 0:01:22.036 having been one of the people who helped come up 0:01:22.036 --> 0:01:23.876 with the idea for it in the first place, and 0:01:23.916 --> 0:01:27.596 who worked with Facebook to bring it into reality. For 0:01:27.676 --> 0:01:30.596 full disclosure, I also want listeners to know, if you 0:01:30.596 --> 0:01:33.996 don't already, that I continue to advise Facebook on free 0:01:34.076 --> 0:01:37.756 expression related issues in order to discuss the business of 0:01:37.756 --> 0:01:40.876 the Oversight Board. However, I am not the right person 0:01:40.916 --> 0:01:43.036 to do the talking because I am not on the 0:01:43.076 --> 0:01:46.316 Oversight Board, and the Oversight Board is its own independent 0:01:46.356 --> 0:01:50.436 institution for that purpose. We're very fortunate to be welcomed 0:01:50.476 --> 0:01:54.516 on today's show by Jamal Green. Jamal is the Dwight 0:01:54.636 --> 0:01:58.196 Professor of Law at Columbia Law School. He's an expert 0:01:58.236 --> 0:02:01.836 on constitutional decision making and the author of an important 0:02:01.836 --> 0:02:05.276 book called How Rights Went Wrong. He's one of the 0:02:05.316 --> 0:02:09.236 co chairs of the Oversight Board, in which position he 0:02:09.276 --> 0:02:12.716 has a crucial insider's view of the work of the board, 0:02:13.036 --> 0:02:19.716 its purposes, its functions, its limitations, and the challenges it faces. Jamal, 0:02:19.876 --> 0:02:27.076 thank you so much for being Jamal. Let's start with 0:02:27.676 --> 0:02:30.236 the Oversight Board, of which you're one of the chairs, 0:02:30.956 --> 0:02:34.476 and tell me how you decided to take that job 0:02:34.556 --> 0:02:37.476 when it was offered to oh gosh, when the team 0:02:37.476 --> 0:02:41.556 at Facebook that was making the initial choices reached out 0:02:41.556 --> 0:02:44.676 and kind of explained the idea, it sounded like an 0:02:44.716 --> 0:02:47.956 opportunity both to try to make the world slightly better 0:02:48.076 --> 0:02:52.396 and also something that aligned with my own personal interests 0:02:52.396 --> 0:02:55.596 and my professional interests in One of the things I 0:02:55.636 --> 0:02:57.516 do in my day job is think a lot about 0:02:58.276 --> 0:03:00.676 how to balance and optimize different kinds of rights and 0:03:00.716 --> 0:03:03.516 how to think about how rights interact with different kinds 0:03:03.556 --> 0:03:06.476 of institutions. So it had a professional interest for me, 0:03:06.796 --> 0:03:10.396 and given how important what Facebook does in terms of 0:03:10.476 --> 0:03:14.156 content moderation, is seemed like an opportunity to contribute to 0:03:14.196 --> 0:03:15.876 making the world a little bit better. So that's why 0:03:15.916 --> 0:03:21.996 I decided to do it. And he regrets, No, no regrets. Challenges, 0:03:22.076 --> 0:03:23.996 of course, but I knew there would be challenges, so 0:03:23.996 --> 0:03:26.276 I wouldn't say there are any regrets. No good. I'm 0:03:26.316 --> 0:03:29.236 happy to hear that, since my own involvement with the 0:03:29.276 --> 0:03:31.716 Oversight Board went back to the point where there weren't 0:03:31.716 --> 0:03:33.676 any chairs. In fact, there was only the idea for 0:03:33.716 --> 0:03:36.276 the thing. I want to talk about your work on 0:03:36.436 --> 0:03:39.076 rights and its relationship to some of the themes at 0:03:39.076 --> 0:03:41.756 the Facebook Oversight Board faces. But I want to first 0:03:41.756 --> 0:03:46.356 start with transparency, which is, on the one hand, one 0:03:46.396 --> 0:03:48.956 of the core rationales for the Oversight Board in the 0:03:49.076 --> 0:03:54.316 first place. It exists in some sense to push Facebook 0:03:54.396 --> 0:03:57.196 to be more transparent it's decision making and to make 0:03:57.236 --> 0:04:00.236 its own decisions in a way that is transparent in 0:04:00.236 --> 0:04:05.316 the sense of revealing reasoning and logic. Transparency, though, has 0:04:05.396 --> 0:04:08.556 also been a great challenge for Facebook to put it 0:04:08.836 --> 0:04:11.676 some what overly politely, I mean, the company has been 0:04:12.476 --> 0:04:16.756 very badly buffeted by a whole series of leaks which 0:04:17.036 --> 0:04:21.636 have gone to major newspapers, to Congress, to probably the 0:04:21.636 --> 0:04:27.276 Federal Trade Commission, all of which seem primarily structured, not solely, 0:04:27.316 --> 0:04:30.476 but primarily structure around this idea that the company hasn't 0:04:30.476 --> 0:04:34.916 been transparent enough in aspects of its decision making. So 0:04:34.956 --> 0:04:38.716 I'm wondering what you think the Oversight Board can do 0:04:39.436 --> 0:04:43.756 to push Facebook towards greater transparency that hasn't already been 0:04:43.796 --> 0:04:47.676 done by this leaking process. Well, I think there's a 0:04:47.676 --> 0:04:50.516 couple of different ways in which one can talk about transparency, 0:04:50.556 --> 0:04:55.916 and Facebook has that issue along several dimensions. The first 0:04:55.916 --> 0:04:59.156 thing to say is, you know, we care about transparency 0:04:59.916 --> 0:05:04.156 to the degree that the particular activity touches on lots 0:05:04.156 --> 0:05:07.676 and lots of people in a way that one thinks 0:05:07.716 --> 0:05:11.356 that a company or or institution needs to be held 0:05:11.436 --> 0:05:13.996 accountable in some way. Right, So if Facebook had no 0:05:14.036 --> 0:05:16.196 reach at all, it was just a sort of private 0:05:16.236 --> 0:05:19.436 company making its own decisions, we wouldn't care that much 0:05:19.436 --> 0:05:21.916 about transparency, just as we don't care that much about 0:05:21.956 --> 0:05:24.916 transparency for you know, the person who makes our kids 0:05:24.916 --> 0:05:28.756 toys or something. But it's because Facebook's reach is so 0:05:28.836 --> 0:05:32.196 broad and now in a global sense, so broad and 0:05:32.356 --> 0:05:36.556 touches on a basic social policy when it comes to vaccines, 0:05:36.676 --> 0:05:39.196 or it comes to political elections and all sorts of 0:05:39.196 --> 0:05:42.796 things that Facebook policies can have an impact on. That's 0:05:42.796 --> 0:05:45.276 where the demand for transparency comes in. It's in a 0:05:45.316 --> 0:05:48.996 sense proportional to the reach of the company. And so 0:05:49.036 --> 0:05:51.636 when we say something like Facebook has not been transparent enough, 0:05:52.036 --> 0:05:54.316 I think the right way to understand that is to 0:05:54.356 --> 0:05:56.916 say that its transparency is not on par with its reach. 0:05:57.356 --> 0:05:59.596 And I think that's a legitimate problem for the company. 0:06:00.196 --> 0:06:03.236 We think about leaks, right, A lot of the leaking 0:06:03.436 --> 0:06:08.276 has been related to internal research that Facebook is conducted 0:06:08.276 --> 0:06:11.556 and the degree to which responded to that internal research. 0:06:11.676 --> 0:06:15.396 And that's a certain kind of transparency problem that I 0:06:15.436 --> 0:06:18.076 think the board is not centrally focused on. Although the 0:06:18.116 --> 0:06:21.436 board is certainly interested in what the company's doing. The 0:06:21.476 --> 0:06:24.276 ways in which what the board does can be most 0:06:24.276 --> 0:06:28.076 directly relevant to Facebook's transparency is it makes lots and 0:06:28.116 --> 0:06:33.076 lots of decisions without fully explaining why it's making those decisions. Right, So, 0:06:33.076 --> 0:06:36.556 when it comes to content people's content gets taken down, 0:06:36.596 --> 0:06:39.196 other content gets left up. There's not much of an 0:06:39.196 --> 0:06:42.996 opinion written about it. Facebook barely gives reasons for why 0:06:43.036 --> 0:06:45.156 it does what it does. It seems like it's acting 0:06:45.156 --> 0:06:48.396 in consistently in a variety of ways. And so what 0:06:48.436 --> 0:06:51.316 the board can do is open up that decision making 0:06:51.316 --> 0:06:54.436 process to show what kinds of trade offs are being made, 0:06:54.796 --> 0:06:58.756 to make recommendations and in some cases of binding decisions 0:06:58.836 --> 0:07:01.436 on how those tradeoffs should be made, and doing it 0:07:01.436 --> 0:07:04.636 in a way that we write opinions, we publish those opinions, 0:07:04.636 --> 0:07:07.636 We tell people exactly what our sources of information are, 0:07:07.676 --> 0:07:10.396 what exactly we're weighing, what the kind of governing I'll 0:07:10.396 --> 0:07:13.476 put this in quotes law is. So what are the 0:07:13.516 --> 0:07:16.556 resources for relying upon to make those decisions in a 0:07:16.596 --> 0:07:19.076 way the company just hasn't And I think that that 0:07:19.196 --> 0:07:22.996 goes directly to how fairly the company treats its choosers. Again, 0:07:23.036 --> 0:07:26.276 there's a broader question of the transparency of any institution 0:07:26.356 --> 0:07:29.196 that's a wielding this much power. What the board does 0:07:29.356 --> 0:07:31.716 is one element of that. So I both want to 0:07:31.796 --> 0:07:34.516 explore the distinction that you're offering jamal between two different 0:07:34.596 --> 0:07:38.276 kinds of transparency and then also maybe ask you whether 0:07:38.316 --> 0:07:40.596 they might have more in common than you're suggesting. So 0:07:40.636 --> 0:07:43.596 I hear you saying that what the oversight board does best, 0:07:43.596 --> 0:07:47.716 and I agree with this is focusing on let's call 0:07:47.756 --> 0:07:51.076 it reason giving transparency, right, This is the idea that 0:07:51.116 --> 0:07:54.116 when the company makes a decision about some piece of content, 0:07:54.476 --> 0:07:57.196 it needs to follow principles and it needs to reveal 0:07:57.556 --> 0:08:00.796 transparently what its reasoning process was, and often they don't. 0:08:01.116 --> 0:08:03.876 And since the oversight board itself exercises that kind of 0:08:03.876 --> 0:08:07.956 transparency in making its own decisions, because you guys, tell 0:08:07.996 --> 0:08:09.996 everybody here or the principle is that we invoked, here 0:08:09.996 --> 0:08:12.876 are the facts we relied on, here are the moral 0:08:12.916 --> 0:08:14.996 ideas that we're relevant to our analysis, and here's the 0:08:15.076 --> 0:08:17.476 kind of conclusion that we reached after weighing these factors. 0:08:18.196 --> 0:08:21.116 You can urge Facebook to do something similar. So I 0:08:21.156 --> 0:08:23.636 get that side of the distinction. And then I hear 0:08:23.676 --> 0:08:26.316 you suggesting that there's also a kind of transparency of 0:08:26.836 --> 0:08:28.876 you know, how much does a company disclose about its 0:08:28.916 --> 0:08:33.676 internal research about the consequences of its actions? And so 0:08:33.796 --> 0:08:35.516 first let's just see if that is the distinction you're 0:08:35.956 --> 0:08:38.996 pushing here. So, yes, that is the distinction I'm pushing. 0:08:39.156 --> 0:08:41.956 I'll be curious what the follow up question is, because 0:08:41.956 --> 0:08:44.716 I do think that there is a relation between them, 0:08:44.716 --> 0:08:48.596 but I'll let you Yeah, I mean to be collaborative. 0:08:48.636 --> 0:08:51.716 I mean, I think that those are in some sense different. 0:08:51.756 --> 0:08:54.956 But if you think about, you know, Facebook having access 0:08:55.036 --> 0:08:59.716 to internal research that suggested maybe some of its policies 0:08:59.756 --> 0:09:03.076 we're having a negative effect on users, I think the 0:09:03.116 --> 0:09:05.996 public call is not merely that that fact should have 0:09:05.996 --> 0:09:08.956 been known, but that people want to know that Facebook's 0:09:08.876 --> 0:09:13.156 decision making process where it decided to continue these services, 0:09:13.236 --> 0:09:15.396 or that it decided to tweak them in certain ways 0:09:15.916 --> 0:09:19.156 rather than eliminate them all together, was a reasoning process 0:09:19.196 --> 0:09:22.236 that they can hear. So I mean, in that sense, 0:09:22.756 --> 0:09:25.956 the decision making is always you take certain facts, you 0:09:25.996 --> 0:09:29.596 have certain values, and you try to bring those together 0:09:29.836 --> 0:09:32.796 and make a decision, and then if you're being transparent 0:09:32.836 --> 0:09:35.076 about your decision making, you tell people why you did that. 0:09:35.556 --> 0:09:37.876 And so I actually have the ends think that that's 0:09:37.876 --> 0:09:40.796 actually more similar to the kind of transparency and decision 0:09:40.796 --> 0:09:43.356 making that you're talking about. I don't think it's just 0:09:43.396 --> 0:09:44.796 that people were saying, well, they should have told us 0:09:44.836 --> 0:09:47.036 about this research. I think they're saying they should have 0:09:47.036 --> 0:09:50.516 taken this research into account in making their decisions. And 0:09:50.556 --> 0:09:52.476 when Facebook says, which is sort of all they've said 0:09:52.516 --> 0:09:55.836 so far, oh, don't worry, we did, people's response to 0:09:55.836 --> 0:09:57.236 that is to say, well, how do we know you 0:09:57.276 --> 0:10:00.076 did that right? And had they been more transparent about 0:10:00.116 --> 0:10:02.756 their decision making process, I think they could have said, well, 0:10:02.756 --> 0:10:04.716 this is how we took that into account, or this 0:10:04.796 --> 0:10:06.596 is how we didn't take it into account. I do 0:10:06.636 --> 0:10:10.076 think it's the case, certainly that part of what the 0:10:10.116 --> 0:10:13.716 board does when it sees a decision that's made by 0:10:13.716 --> 0:10:16.956 Facebook or increasingly the board has weighed in and will 0:10:16.996 --> 0:10:21.276 weigh in on actual policies that Facebook is implementing and considering. 0:10:21.836 --> 0:10:25.876 One of the things we care about is why did 0:10:25.916 --> 0:10:30.436 the mistake happen? And when we say a mistake, we 0:10:30.756 --> 0:10:33.956 use a few different resources figuring out what counts as 0:10:33.956 --> 0:10:36.676 a mistake. Right, So we think about whether Facebook as 0:10:36.716 --> 0:10:41.356 apply its community standards accurately. We think about if Facebook 0:10:41.476 --> 0:10:45.196 is acting consistent with how it understands its values, and 0:10:45.316 --> 0:10:47.876 we think about international human rights law, which and the 0:10:47.956 --> 0:10:51.676 norms associated with the international human rights system to which 0:10:51.716 --> 0:10:55.916 Facebook has committed itself. It may well be that there 0:10:55.956 --> 0:11:00.476 are decisions that Facebook makes that are inconsistent with those 0:11:00.596 --> 0:11:03.556 values and norms. And one of the things that the 0:11:03.596 --> 0:11:05.676 Board tries to do and has been trying to do 0:11:05.756 --> 0:11:07.836 in decision is not just make an up or down 0:11:07.876 --> 0:11:10.036 decision on a piece of content. I'd say why with 0:11:10.076 --> 0:11:12.916 this mistake made? And in order to do that, it 0:11:12.956 --> 0:11:14.996 may sometimes be the case, right that we need to 0:11:14.996 --> 0:11:17.716 know more about what inputs there were into particular kinds 0:11:17.716 --> 0:11:20.916 of content decisions. And I think the Board in asking 0:11:20.916 --> 0:11:24.876 Facebook questions about that, in exposing what answers it gives 0:11:24.876 --> 0:11:27.156 to us when it refuses to give answers to us, 0:11:27.236 --> 0:11:31.316 in pushing Facebook to be more clear about what influences 0:11:31.356 --> 0:11:33.196 it has when it's decision making. Right. So, one of 0:11:33.196 --> 0:11:36.036 the recommendations the Board has made that Facebook has taken 0:11:36.116 --> 0:11:39.676 up is to be more clear about when governments make 0:11:39.716 --> 0:11:43.436 requests or Facebook to remove content. Right. So that's relevant 0:11:43.476 --> 0:11:45.876 to transparency as well. Right, So, there are things the 0:11:45.916 --> 0:11:49.516 Board can do that get at that deeper transparency. What 0:11:49.556 --> 0:11:52.356 I meant to say is that there are some very 0:11:52.396 --> 0:11:57.476 deep issues about who governs us that at the moment, 0:11:57.916 --> 0:12:01.796 the Board's jurisdiction and scope are limited. There's no question 0:12:01.836 --> 0:12:04.476 about that. There's very important progress that the Board can 0:12:04.476 --> 0:12:07.876 and does make on I'm trying to get Facebook to 0:12:07.916 --> 0:12:10.676 treat its users better or more consistently, to start trying 0:12:10.676 --> 0:12:14.036 to get at what's influencing its decisions. But there is 0:12:14.076 --> 0:12:16.996 always going to be a deeper question about the private 0:12:17.036 --> 0:12:22.156 companies engaged in far reaching activities that I think our 0:12:22.236 --> 0:12:24.996 questions for sort of all of society, Lots of different 0:12:25.076 --> 0:12:30.396 kinds of institutions, the Board, but also whistleblowers and journalists 0:12:30.436 --> 0:12:34.516 and researchers and civil society organizations that are equally if 0:12:34.556 --> 0:12:36.956 not more situated to get into than the Board itself. 0:12:38.436 --> 0:12:40.236 Let me ask you about one point where there was 0:12:40.276 --> 0:12:43.596 some overlap between what the Board said in its recent 0:12:43.596 --> 0:12:47.836 transparency reports and what the whistleblowers materials disclosed, and that 0:12:47.916 --> 0:12:52.036 was the program sometimes called cross Check, through which Facebook 0:12:52.036 --> 0:12:57.236 initially was trying to address just a relatively smaller number 0:12:57.396 --> 0:13:02.956 of distinctive users with respect to their newsworthiness of what 0:13:03.036 --> 0:13:05.116 was being posted on the platform, but that extended to 0:13:05.156 --> 0:13:10.516 cover really a very large number of users, many more 0:13:10.556 --> 0:13:13.876 than Facebook had acknowledged, and it seems many more than 0:13:13.996 --> 0:13:17.356 Facebook told the Oversight Board when the Oversight Board asked 0:13:17.396 --> 0:13:20.156 point blank about this in the course of the Trump 0:13:20.196 --> 0:13:23.116 deplatforming decision, what can you say about that and what 0:13:23.196 --> 0:13:26.756 do the oversight boards say about that recently? So I 0:13:26.796 --> 0:13:29.436 can say a bit about the sort of background here, 0:13:29.676 --> 0:13:32.596 but I'll note that Facebook has given to the board 0:13:33.316 --> 0:13:36.556 a policy advisory requests on how to structure it's cross 0:13:36.636 --> 0:13:39.676 check program. And I wouldn't want to say too much 0:13:39.716 --> 0:13:42.916 in advance of deliberating about that and getting more information 0:13:42.956 --> 0:13:46.036 about it about exactly what problems there might be with 0:13:46.076 --> 0:13:48.916 crosscheck or what the right way to resolve those problems 0:13:48.996 --> 0:13:52.076 might be. But crosscheck is this system that Facebook has 0:13:52.076 --> 0:13:55.916 in place, or had in place, in which it exposes 0:13:55.956 --> 0:14:00.316 certain users to additional layers of review, ostensibly on the 0:14:00.316 --> 0:14:03.236 theory that they don't want mistakes to be made with 0:14:03.316 --> 0:14:06.196 respect to certain users. And the board asked in the 0:14:06.196 --> 0:14:09.556 Trump decision about this program, and Facebook it that it 0:14:09.596 --> 0:14:12.276 was it was only used for a small number of users, 0:14:12.356 --> 0:14:14.156 and it turns out to be it's a few million. 0:14:14.396 --> 0:14:17.516 Facebook said to the board later that that was a 0:14:17.556 --> 0:14:20.116 small number in relation to the number of people on Facebook, 0:14:20.116 --> 0:14:22.316 which is of course true. But it's true in a 0:14:22.396 --> 0:14:25.356 kind of layally way. So what the board said and 0:14:25.356 --> 0:14:27.676 it's recent, I can say, since we both law professors, 0:14:27.676 --> 0:14:29.276 I can say that you're using the word layally in 0:14:29.276 --> 0:14:32.156 the negative sense of that term. That's that's right, right. So, 0:14:32.436 --> 0:14:35.356 if you're dealing with someone in an adversarial posture, as 0:14:35.436 --> 0:14:38.076 lawyers often are, right, sometimes if they ask you a 0:14:38.156 --> 0:14:40.716 question you answer it in the most narrow possible way. 0:14:40.796 --> 0:14:43.036 You might still be truthful, but in some ways it's 0:14:43.036 --> 0:14:46.156 misleading if you're being very narrow about it. But if 0:14:46.156 --> 0:14:47.636 you're you know, talking to a friend of yours and 0:14:47.716 --> 0:14:50.916 they ask you about some piece of information, it would 0:14:50.916 --> 0:14:53.996 be strange to be excessively narrow about that. And I 0:14:54.396 --> 0:14:57.716 what I would say is Facebook, in not being I 0:14:57.756 --> 0:15:02.116 think fully forthcoming, I think treated this in two adversarial 0:15:02.236 --> 0:15:04.876 sort of way. Right. So when we ask them for information, 0:15:05.356 --> 0:15:08.076 we think that they should give us the full context 0:15:08.196 --> 0:15:10.756 and try to be as helpful as they can in 0:15:10.796 --> 0:15:14.756 providing the board with information. And we told them as much, right. So, 0:15:15.116 --> 0:15:17.476 going forward, Facebook has promised that it's going to be 0:15:17.556 --> 0:15:20.356 more contextual in the way in which it responds to 0:15:20.356 --> 0:15:23.276 information requests, and that's going to be I think very 0:15:23.316 --> 0:15:25.876 helpful for the board to try to do. It's not 0:15:26.076 --> 0:15:29.116 better because you don't always know what you don't know right, 0:15:29.156 --> 0:15:33.276 and so understanding better exactly how the cross check program 0:15:33.316 --> 0:15:36.716 works can be helpful in deciding whether it's being applied 0:15:36.756 --> 0:15:40.156 fairly in a particular case, Jamal, I want to turn 0:15:40.236 --> 0:15:43.156 to the core of the Oversight Board's job, which is 0:15:43.236 --> 0:15:47.316 decision making, and here I'll be really curious to hear 0:15:47.356 --> 0:15:52.156 from you about how your distinctive approach to decision making 0:15:52.996 --> 0:15:56.196 plays out. You published an amazing book this year called 0:15:56.316 --> 0:15:59.116 How Rights Went Wrong, and that book, in turn drew 0:15:59.156 --> 0:16:02.756 on some of your earlier scholarship, which I and others 0:16:02.796 --> 0:16:05.316 read in the course of trying to think about how 0:16:05.316 --> 0:16:07.716 the Oversight Board should make its decisions in the first place. 0:16:07.796 --> 0:16:09.996 So some of your approaches I think maybe you already 0:16:09.996 --> 0:16:12.276 baked in before you got there. But I wonder if 0:16:12.276 --> 0:16:14.876 you would start by just saying something about your distinctive 0:16:14.956 --> 0:16:19.116 view of how courts or bodies that are sort of 0:16:19.156 --> 0:16:22.956 like chords, like the Oversight Board, should decide cases where 0:16:22.956 --> 0:16:27.196 there are reasonable arguments on both sides. Sure, and I'm 0:16:27.196 --> 0:16:30.236 happy to talk about this with a couple of caveats right, 0:16:30.316 --> 0:16:34.796 one being that it's not just my approach, right, I 0:16:34.836 --> 0:16:38.276 think I have a particular angle on it, But it 0:16:38.396 --> 0:16:41.316 is a pushback against the way in which US courts 0:16:41.316 --> 0:16:44.236 and often US thinkers about rights tend to think about rights, 0:16:44.556 --> 0:16:48.156 less of a pushback against some global standards. And the 0:16:48.196 --> 0:16:51.276 second thing I'll say is we are collaborative board, right, 0:16:51.316 --> 0:16:53.516 And so if I were writing all of the opinions 0:16:53.556 --> 0:16:55.436 just by myself, they might look a little bit different 0:16:55.436 --> 0:16:59.516 than writing opinions when twenty members have to more or 0:16:59.596 --> 0:17:02.876 less agree on them. The general point is that when 0:17:02.916 --> 0:17:07.076 we're talking about rights conflicts, that rights conflicts are very often, 0:17:07.196 --> 0:17:12.076 not always, but very often conflicts in which people have 0:17:12.196 --> 0:17:15.596 reasonable disagreement about how to apply a set of more 0:17:15.676 --> 0:17:18.596 or less shared values at a high level of generality, 0:17:18.956 --> 0:17:21.396 but they disagree on how to apply them in the 0:17:21.396 --> 0:17:25.356 particular case. When we're in that situation, it's not that 0:17:25.516 --> 0:17:29.756 helpful to pick out just a few rights that we 0:17:30.236 --> 0:17:33.596 think are important and essentialize them so that they are 0:17:33.596 --> 0:17:37.276 applied kind of absolutely whenever they're because that will tend 0:17:37.316 --> 0:17:42.796 to silence one or the other side of these rights conflicts. 0:17:42.796 --> 0:17:44.596 And I think in the US we tend to do 0:17:44.636 --> 0:17:49.396 this with the First Amendment. So the moment someone invokes speech, 0:17:49.636 --> 0:17:53.076 there's a battle over saying who has the speech right 0:17:53.276 --> 0:17:55.716 or whether there's a speech right or not. And because 0:17:55.756 --> 0:17:57.956 we know the stakes of that battle are extremely high, 0:17:57.996 --> 0:17:59.876 that you just sort of win if you get to 0:17:59.916 --> 0:18:02.476 say that you have a speech right. And what I 0:18:02.516 --> 0:18:05.796 try to urge in the book is in freedom of 0:18:06.196 --> 0:18:10.876 speech cases, as in many others, that a tremendous institutional 0:18:10.956 --> 0:18:14.036 variation in the ways in which speech might be affected. Right, 0:18:14.076 --> 0:18:17.756 So a purge of all people who are opposed to 0:18:17.756 --> 0:18:20.276 the government in which you put them in prison is 0:18:20.396 --> 0:18:24.396 extremely different from let's say, a university deciding how to 0:18:24.396 --> 0:18:27.956 regulate the speech of its students, or a platform deciding 0:18:28.236 --> 0:18:30.476 who is going to be able to amplify their content 0:18:30.796 --> 0:18:33.916 and spread it around the world. All forms of regulation 0:18:33.956 --> 0:18:36.636 of speech. But they're in very very different contexts, and 0:18:36.676 --> 0:18:39.236 that in some of those contexts we have to think 0:18:39.316 --> 0:18:42.916 more carefully about the various other values that we think 0:18:42.956 --> 0:18:46.116 are important than we do another of those contexts. Right, So, 0:18:46.516 --> 0:18:49.716 we care about national security, but we don't care about 0:18:49.716 --> 0:18:52.916 it so much that we allow purges of our political enemies. 0:18:53.356 --> 0:18:56.716 But the fact that we care about hate speech or 0:18:56.756 --> 0:19:01.036 we care about amplification of misinformation? Are values that we 0:19:01.116 --> 0:19:03.476 might think are sufficiently important to put some kinds of 0:19:03.476 --> 0:19:06.116 restrictions on who has access to certain kinds of platforms, 0:19:06.396 --> 0:19:08.636 And so even though they're both speech cases, they're all 0:19:08.636 --> 0:19:10.436 speech case. Is we have to think carefully about what's 0:19:10.436 --> 0:19:12.596 on the other side of the balance, depending on the 0:19:12.596 --> 0:19:16.396 institutional context. Can I ask you a philosophical question around that, 0:19:16.476 --> 0:19:19.516 Jamal that I find myself struggling with very much right now, 0:19:19.956 --> 0:19:22.196 and I don't think there's a simple answer to it. 0:19:22.196 --> 0:19:27.076 It has to do with the boundary between misinformation on 0:19:27.116 --> 0:19:29.756 matters that you and I would probably agree have a 0:19:29.836 --> 0:19:35.396 fact associated with them, and misinformation on matters that have 0:19:35.556 --> 0:19:40.236 a fact associated with it but has become so politicized 0:19:40.796 --> 0:19:44.636 that it becomes a stand in for somebody's political beliefs 0:19:44.676 --> 0:19:47.596 and values. Maybe a climate change would be a good example. 0:19:48.156 --> 0:19:49.636 I think you and I both think that there is 0:19:49.676 --> 0:19:52.276 a science of it, and the scientists are doing their 0:19:52.276 --> 0:19:53.636 best to get at it. They might be right, they 0:19:53.716 --> 0:19:56.716 might be wrong, but they achieve consensus they follow their process. 0:19:57.636 --> 0:20:01.516 But when people argue about climate change and whether it's 0:20:01.516 --> 0:20:04.116 man made, a lot of people are using in that argument, 0:20:04.156 --> 0:20:07.116 which is nominally an argument about facts, as a kind 0:20:07.156 --> 0:20:10.916 of stand in for their political points of view. And 0:20:11.636 --> 0:20:16.156 once that happens, the differences in what people are saying 0:20:16.236 --> 0:20:19.676 could be put into the box of misinformation if we're 0:20:19.716 --> 0:20:22.676 confident that we know what the science says, and in 0:20:22.676 --> 0:20:24.676 that case, maybe it's not so important to preserve different 0:20:24.676 --> 0:20:26.916 points of view, or they could be put into the 0:20:26.956 --> 0:20:30.836 category of political argument about political identity and about what 0:20:30.876 --> 0:20:33.796 should be done in the future, and that's really important 0:20:34.036 --> 0:20:37.956 and would probably deserve a lot more protection. So I 0:20:37.956 --> 0:20:40.436 deliberately didn't choose COVID because it's too close to home 0:20:40.476 --> 0:20:43.076 and too controversial, But climate change is still pretty darn important. 0:20:43.716 --> 0:20:46.196 So I guess I'm wondering what you think about that. 0:20:46.236 --> 0:20:47.956 And again, it's not that I think there's a particularly 0:20:48.076 --> 0:20:51.276 right answer to it. I just think it's a hard problem. Yeah, no, don't. 0:20:51.356 --> 0:20:53.436 I think it's definitely a hard problem. I don't think 0:20:53.476 --> 0:20:57.716 it has a sort of abstract answer. What I would 0:20:57.756 --> 0:21:02.076 say is a couple of things. So one is the 0:21:02.476 --> 0:21:07.596 bare fact that something is false perhaps should not engage 0:21:07.676 --> 0:21:11.876 or enrage or site us as much as whether something 0:21:11.876 --> 0:21:14.796 that is false is leading people to do something that 0:21:14.916 --> 0:21:18.476 is more materially harmful, So that just goes back into 0:21:18.556 --> 0:21:22.116 sort of COVID misinformation might be an exlightly different category 0:21:22.156 --> 0:21:25.676 than something like climate change, where the latter is might 0:21:25.676 --> 0:21:28.116 be influencing policy in some way, but in a somewhat 0:21:28.156 --> 0:21:32.356 indirect way. Whereas if someone really does think that if 0:21:32.356 --> 0:21:36.036 they inject bleach into their veins, or if they take 0:21:36.236 --> 0:21:39.396 some off labeled drug that might hurt them, that's in 0:21:39.436 --> 0:21:41.876 a different kind of category in terms of the immediacy 0:21:41.876 --> 0:21:45.156 of harm. And that's important, right because we because it's 0:21:45.236 --> 0:21:48.276 very hard, as you say, for these philosophical reasons to 0:21:48.356 --> 0:21:51.036 sort of adjudicate these things in the abstract, but when 0:21:51.036 --> 0:21:53.836 we are able to connect them to more concrete harms, 0:21:53.836 --> 0:21:56.956 that affects how we feel about regulating them, even if 0:21:56.996 --> 0:22:00.836 we can't resolve the philosophical issue that you just raised. Right, So, 0:22:01.396 --> 0:22:04.116 there are various forms of misinformation and ways in which 0:22:04.156 --> 0:22:08.076 we mislead each other. There's a long spectrum from pure 0:22:08.116 --> 0:22:10.836 truth to pure life, and we're often somewhere in the 0:22:10.836 --> 0:22:13.836 middle of that in our political discourse. So I tend 0:22:13.836 --> 0:22:17.276 to think that really the only productive way that a 0:22:17.356 --> 0:22:20.356 regulator of some kind can respond to that is to 0:22:20.396 --> 0:22:24.676 try to focus on direct and concrete harms. We'll be 0:22:24.756 --> 0:22:38.596 right back, Jamal on your oversight board. There is one 0:22:38.636 --> 0:22:41.716 of your CoA chairs, Michael McConnell's a retired federal judge 0:22:41.756 --> 0:22:44.876 also a law professor. Do people like him, or people 0:22:44.876 --> 0:22:48.756 who come out of one particular system find it relatively 0:22:48.876 --> 0:22:53.076 simple and seamless to shift to a more overtly recognizing 0:22:53.196 --> 0:22:55.116