WEBVTT 00:00.000 --> 00:07.340 And thanks, everybody, for having us here to talk about some of our favorite subjects, online speech and privacy. 00:07.560 --> 00:22.860 I'm Cathy Gellis, half of this duo, and I'm going to be going first because if we're going to be talking about potential changes to Section 230, which is what the EARN IT Act does in part, we should have a good idea about what Section 230 is, how it works, 00:22.860 --> 00:27.940 and really understand it before we break it and find out that maybe that was a mistake. 00:27.940 --> 00:28.960 It will be a mistake. 00:28.960 --> 00:29.460 Stay tuned. 00:29.460 --> 00:30.800 We will explain why. 00:31.140 --> 00:36.320 By way of introduction, I am a solo cyber lawyer in the San Francisco Bay Area. 00:36.380 --> 00:41.640 I like to work on online speech issues, including Section 230 issues. 00:41.660 --> 00:42.640 I am a lawyer. 00:42.640 --> 00:43.800 Rihanna's a lawyer. 00:43.800 --> 00:45.300 We are not your lawyers. 00:45.340 --> 00:49.560 If we were your lawyers, we would have all signed letters saying so, and there's been no letters. 00:49.560 --> 00:50.960 So just to make that clear. 00:50.960 --> 00:53.240 But we do know what we're talking about. 00:53.360 --> 00:55.840 So this presentation has two parts. 00:55.840 --> 01:00.840 The reason I'm going first, doing the crash course, and then I will be joined by Rihanna to talk more. 02:58.300 --> 02:59.620 Hi, everyone. 02:59.620 --> 03:03.000 We are back again with more technical difficulties. 03:03.000 --> 03:05.100 But have no fear, I've got everyone back. 03:05.100 --> 03:08.580 I don't know if anyone's ever done OBS over a virtual desktop. 03:08.660 --> 03:26.660 Apparently, tip for the other folks learning from this stream, if you are logged into a Zoom in a virtual machine using OBS, and somebody else logs into that virtual machine, Zoom disconnects the user that is connected to Zoom on that virtual machine. 03:26.660 --> 03:27.460 So we're back. 03:27.460 --> 03:30.660 We've learned this new thing about technology. 03:30.700 --> 03:35.400 I will give it back over to Kathy, who's going to talk about the overarching principles. 03:37.380 --> 03:38.720 Okay, thanks, Whitney. 03:38.740 --> 03:40.760 I'm not quite sure where we got cut off. 03:40.760 --> 03:46.260 But Rihanna and I, both lawyers, not your lawyers, but it does mean we know what we're talking about. 03:46.760 --> 03:50.680 The first bit that we're going to do is crash course in Section 230. 03:50.680 --> 04:07.200 Because speaking of learning curves, stuff breaks, but it'd be really stupid to break it on purpose, because we didn't understand this really great and important law before we start changing it with proposals like the earn it and various other things that are floating through the halls of our government. 04:07.520 --> 04:17.060 So getting into it, as we go into this lesson, this crash course, couple of things to keep in mind, overarching principles. 04:17.200 --> 04:21.560 The first thing, Section 230 is a really hot topic these days. 04:21.680 --> 04:23.020 Everybody's talking about it. 04:23.020 --> 04:24.800 Everybody's complaining about it. 04:24.800 --> 04:35.160 Bear in mind that much of what people are complaining about, because people are saying bad stuff, saying and doing bad stuff on the internet, is really a byproduct of we've got a First Amendment. 04:35.420 --> 04:40.180 Freedom of speech tends to protect an awful lot of stuff that we are upset about. 04:40.180 --> 04:48.060 And if we were going to try to change the law to change it, that means we'd have to get rid of the First Amendment, and that's probably not a really good idea. 04:48.720 --> 04:56.960 So it also means that basically, if we're unhappy about something, the correct policy choice is not to change Section 230, because it's not actually Section 230's fault. 04:57.540 --> 05:09.140 A second overarching point that we need to bear in mind is that litigation can be cripplingly expensive, even when the position that you've got as a defendant is the better one. 05:09.140 --> 05:13.420 The fact you might win the lawsuit doesn't necessarily help. 05:13.420 --> 05:28.840 Lawsuits are expensive and potentially bankrupting, and what Section 230 is designed to do is to make sure that platforms can stay in business without feeling like they're going to be financially obliterated because one of their users says or does something stupid, 05:28.840 --> 05:31.400 and somebody tries to hold them responsible to it. 05:31.400 --> 05:41.840 Because if you think about the tremendous volume of what even the smallest platforms end up facilitating in terms of user-generated content, that's enormous. 05:41.840 --> 05:46.860 Even if they just had to, like, call up a lawyer to find out, do we need to really respond to this? 05:46.860 --> 05:48.180 That gets expensive. 05:48.180 --> 05:49.420 That meter ticks. 05:49.420 --> 05:50.980 It's not realistic. 05:50.980 --> 06:05.320 And what Section 230 is designed to do is make sure that a platform cannot be bled dry financially by getting called into court by legal threats, real or even actual ones as well. 06:05.920 --> 06:15.760 Put another way, the overall goal of Section 230, and we're going to get back to this a lot, is to ensure that we get the most good stuff online and the least bad. 06:16.080 --> 06:21.520 There are a lot of really cool things that people say or do over the internet. 06:21.520 --> 06:26.700 We want to make sure that the internet remains as available as possible to get the most of that. 06:26.780 --> 06:34.120 If platforms had to fear liability, they would not be in the position to facilitate all that good stuff. 06:34.120 --> 06:37.720 It would end up much more limited or maybe not be possible at all. 06:37.820 --> 06:42.960 At the same time, there's a lot of crap on the internet, and we would like to minimize the crap. 06:43.120 --> 06:53.280 So the other point of Section 230 is to put platforms in a position where they can minimize the crap without worrying about getting into trouble if they actually do the best they can on that front. 06:53.300 --> 06:58.520 So in order to optimize the most good and least bad, we have Section 230. 06:58.520 --> 07:05.700 It doesn't mean that we'll ever hit optimum on either front, but it gets us as close as we can reasonably get. 07:06.140 --> 07:08.320 And that gets to the fourth point. 07:08.600 --> 07:11.780 Effective regulation doesn't need to be punitive. 07:11.800 --> 07:13.840 Carrots can work best with sticks. 07:13.840 --> 07:17.280 And what Section 230 is best understood as is all carrot, no stick. 07:17.280 --> 07:27.880 That basically, we want the platforms to be there to get us the most good stuff and the least bad stuff, and we're going to make it... we're going to incentivize them to do the best they can on both fronts. 07:27.880 --> 07:32.180 And the reason they're going to be able to do the best they can is because they're not scared. 07:32.180 --> 07:35.960 They're not scared of being punished financially through liability, etc. 07:36.800 --> 07:45.080 That's the magic of Section 230, to make it possible that the platforms and policy can all be aligned to do the best that we can. 07:45.240 --> 07:47.140 That's why we have the Internet today. 07:47.200 --> 07:54.580 So let's take a look at our next slide and talk about how we got Section 230. 07:54.580 --> 07:59.900 So if we get in our time machine and go back to 1995, a couple of things were happening. 07:59.920 --> 08:06.700 The first thing that was happening was that the Internet, which had been existing for quite some time, but was starting to enter the public consciousness. 08:07.280 --> 08:17.000 And as it hit the public consciousness, Congress became aware of it and started to become concerned that, oh my gosh, the Internet, there's porn on the Internet. 08:17.000 --> 08:18.580 Oh no, this is terrible. 08:19.960 --> 08:23.880 Some of it, probably not so great, but some of it is maybe a little overreaction. 08:23.880 --> 08:28.120 But in any case, they were very concerned about the amount of porn on the Internet and didn't want any of it there. 08:28.880 --> 08:35.700 Meanwhile, there was also this lawsuit that came up in a state court of New York involving the system Prodigy. 08:35.700 --> 08:41.620 So Prodigy, at the time, there were a number of dial-up, bulletin board-type services. 08:41.620 --> 08:43.540 Some were somewhat large and commercial. 08:43.640 --> 08:47.000 Prodigy, CompuServe, eventually AOL. 08:47.600 --> 08:54.780 They were small in the sense of not as big as today's platforms, but they were relatively big. 08:54.780 --> 08:58.320 Prodigy had been around for 10 years, at least, at that point. 08:58.320 --> 09:01.140 And they handled an awful lot of user-generated content. 09:01.540 --> 09:05.880 And a user had arguably defamed the plaintiff in this lawsuit. 09:05.880 --> 09:15.660 But the plaintiff decided to Prodigy, basically under the theory of how dare you produce a system that allowed a user to defame me. 09:15.960 --> 09:17.200 And they filed this lawsuit. 09:17.540 --> 09:20.940 And a judge in a state court in New York said, well, that sounds good. 09:20.940 --> 09:22.340 Go ahead with your lawsuit. 09:22.340 --> 09:25.460 Sure, Prodigy should be in trouble for allowing this to happen. 09:25.740 --> 09:29.700 And a couple of members of Congress looked at that case with alarm. 09:29.700 --> 09:32.340 Senator Wyden and then Congressman Chris Cox. 09:32.340 --> 09:34.260 And they said, this is not going to be good. 09:34.260 --> 09:35.520 The Internet is coming. 09:35.520 --> 09:37.000 We've got all these cool services. 09:37.000 --> 09:44.840 And they're going to not be able to be available to provide, to facilitate user-generated speech. 09:44.840 --> 09:47.920 Nor are they going to be in a position to take down the terrible stuff. 09:47.920 --> 09:49.760 So again, we have a crisis. 09:49.760 --> 09:53.780 If this law becomes a thing, we won't get the most good stuff. 09:53.780 --> 09:56.060 We won't be able to minimize the most bad stuff. 09:56.060 --> 09:57.240 This is a problem. 09:57.420 --> 10:01.960 So these two ideas came together and formed the Communications Decency Act of 1996. 10:02.540 --> 10:12.540 And then there was a lawsuit challenging most of it because the getting rid of porn parts happened to infringe on the First Amendment quite severely. 10:12.760 --> 10:14.240 And in a case called Reno v. 10:14.420 --> 10:16.000 ECLU, the U.S. 10:16.000 --> 10:19.240 Supreme Court said, nope, those parts go away. 10:19.300 --> 10:25.100 They made clear that the First Amendment applies to online speech just as much as it applies to offline speech. 10:25.180 --> 10:29.200 Those parts of the CDA went away, but Section 230 survived. 10:29.360 --> 10:36.600 And it survived, and it did its job clearing the way to make sure that people continue to innovate internet services. 10:37.040 --> 10:42.980 So we're now looking at statute. 10:42.980 --> 10:43.720 Don't worry. 10:43.720 --> 10:45.640 If you're not a lawyer, you will survive. 10:45.640 --> 10:46.380 This is okay. 10:46.380 --> 10:47.440 It's a good statute. 10:47.440 --> 10:49.620 It's pretty readable by normal human beings. 10:49.620 --> 10:52.640 But I will give you a scenic tour of some of the parts. 10:52.840 --> 11:03.260 So this particular page that we're on, they put in the beginning of the statute some preamble text, and it basically boils down to Congress finds that the internet is pretty awesome. 11:03.540 --> 11:11.200 They go through two pages worth of awesomeness, and then they get to their policy, and that the policy of the United States is to keep the internet awesome. 11:11.200 --> 11:15.000 So this sort of sets what Congress is trying to accomplish. 11:15.000 --> 11:23.060 And one of the things that it kind of articulates, most good stuff, but it also puts in some, we want to make sure that there's some leverage so we can get rid of the bad stuff. 11:24.880 --> 11:25.680 Here we go. 11:25.680 --> 11:26.640 That's what I just said. 11:26.640 --> 11:27.320 Okay. 11:28.700 --> 11:34.560 This is the most magnificent part of section 230. 11:34.560 --> 11:36.500 It is 26 words. 11:37.080 --> 11:41.240 Professor Jeff Koseff wrote a book called the 26 words that created the internet. 11:41.520 --> 11:48.820 And I've tried to read it out loud at conferences before, and it sounds terrible when pronounced, but it is fantastic. 11:48.820 --> 12:02.680 And if I'm to summarize it in more understandable language, it basically means whoever creates the problematic content is totally responsible for the problematic content, but not the service they used to express it. 12:02.900 --> 12:10.720 And if we look at the bits that I highlighted, the bit at the bottom, the information content provider, that's who created the content. 12:10.720 --> 12:14.940 And that other bit, the interactive computer service, that's the platform. 12:14.940 --> 12:25.860 Internet platforms, interactive computer services, it can be a lot of different things, but it's basically the platform that facilitated a third party's expression. 12:26.740 --> 12:27.940 Here we go. 12:29.240 --> 12:30.100 Okay. 12:30.140 --> 12:35.640 This is the part of the statute that looks at those definitions and basically says what I just told you. 12:35.640 --> 12:42.260 The information content provider is the person who created it, and the interactive computer service is the part that facilitated it. 12:42.260 --> 12:49.120 And that first part basically said the facilitator's not liable, but the people who created it are. 12:49.560 --> 12:56.120 And so there's some question about what sort of, you know, who gets to be counted as an interactive computer service. 12:56.120 --> 12:59.160 And, you know, we think a lot about social media. 12:59.160 --> 13:01.800 That's the easiest thing to kind of get our heads around. 13:01.840 --> 13:19.100 But an email provider, they're facilitating other people's expression, a bulletin board service like Prodigy, comment sections on blogs and online newspapers, these are things that, you know, whoever posted the article is responsible for the content of the article, 13:19.100 --> 13:24.480 but the things that the community provides to comment on the article, that's a textbook case of Section 230. 13:24.720 --> 13:31.380 And then there's also a lot of courts that have looked at this and said, you know, it also could apply to you, individuals. 13:31.380 --> 13:41.360 If you have a Facebook post that takes comments on it, and somebody posts a comment, you're not responsible for the comment if something is wrong with it. 13:41.920 --> 13:43.200 That's an important thing. 13:43.200 --> 13:52.860 I think when people talk about Section 230, they tend to think of it in terms of, well, Google, Twitter, Facebook, and however you feel about these companies tends to color your view. 13:52.860 --> 13:55.220 But understand that it's not about them. 13:55.580 --> 14:00.660 A, yes, true, it was a law that allowed these innovative companies to come and exist. 14:00.780 --> 14:04.400 But it's also what allows any company to come and exist. 14:04.400 --> 14:10.680 And it also is enabling an awful lot of online behavior that ordinary people actually do themselves. 14:12.420 --> 14:16.960 There is a companion piece to Section 230. 14:16.960 --> 14:23.020 It's not something that has actually, like, carried as much weight as the Section C1 part. 14:23.020 --> 14:30.120 But that whole bit of, we want the companies to, we want the platforms to be able to do two things, keep the most good stuff and the least bad stuff. 14:30.120 --> 14:35.600 And so what we've done with that carrot approach is said, okay, here's the carrot. 14:35.600 --> 14:40.260 You're not liable if something wrong gets posted on your systems. 14:40.260 --> 14:44.380 And that's a carrot because it basically means that the companies can say, I'm fine. 14:44.380 --> 14:47.760 I don't have to worry if I'm not able to keep something bad off the internet. 14:47.820 --> 14:53.640 And that means that they're safe to sort of be as expansive as possible to allow the most good stuff. 14:53.700 --> 15:03.040 If they had to fear the punishment , then they would scrutinize a lot of content and take down a lot of content that probably would have been just fine and completely lawful. 15:03.040 --> 15:07.880 So carrot-based approach to make the platforms feel safe to leaving content up. 15:08.060 --> 15:14.840 But we also wanted to make them feel safe to take content down because that's the other part, least bad stuff. 15:14.840 --> 15:22.660 If companies had to fear that they could be punished or held liable for taking content down, they wouldn't bother at all. 15:22.660 --> 15:23.720 It would just be too risky. 15:23.720 --> 15:26.980 But we want them to be partners in helping maintain internet hygiene. 15:26.980 --> 15:30.800 So we've also given them a carrot in this department to say, go ahead. 15:31.760 --> 15:33.620 So here we are next. 15:33.880 --> 15:39.700 The other part of the statute is there's section E where they threw in a couple of other things that end up being really important. 15:39.700 --> 15:42.600 This particular clause is not really important. 15:42.720 --> 15:48.840 They sort of thought back then that there would be a collision with another law that was on books, ECPA. 15:48.920 --> 15:54.080 And so they basically wrote this text to say, no collision, no issue, go away, let's move on. 15:54.080 --> 15:56.160 And so indeed, let's move on. 15:56.760 --> 16:05.440 This is an important clause because if we stop and think about the internet, this is inherently interstate. 16:05.440 --> 16:10.000 It's very difficult to put something on the internet that stays in any one locality. 16:10.000 --> 16:15.720 If you put something on the internet, it's available everywhere, everywhere throughout the country, even throughout the world. 16:16.460 --> 16:23.960 And so we kind of ruin this whole policy if we let each individual state pass a law that will somehow impact it. 16:23.960 --> 16:35.180 So if the companies are feeling safe in Iowa to like, well, we're not going to be liable here, it really doesn't help them if Illinois passes a law that says, no, no, if it happens in Illinois, you're in trouble. 16:35.180 --> 16:37.860 So this is essentially a preemption statute. 16:37.860 --> 16:43.120 And so Congress put this in to say, look, it's federal government, we got it, states sit this one out. 16:43.120 --> 16:46.440 And that's important because internet policy is complex enough. 16:46.440 --> 16:53.100 It really would be overwhelming if all 50 states plus territories and local jurisdictions could have their say. 16:53.860 --> 16:56.440 Okay, but they did put an exception. 16:56.440 --> 17:04.760 So Section 230 applies for almost everything that could be legally wrong with user content with a, but there's a few exceptions. 17:04.760 --> 17:15.020 And one of the big ones that they baked in originally was the thing wrong with the content is that it potentially infringed on an intellectual property right. 17:15.020 --> 17:18.320 Easiest one to get our head around is it might have infringed a copyright. 17:18.540 --> 17:21.480 Then Section 230 doesn't apply. 17:21.480 --> 17:23.460 No protection for the platform. 17:24.180 --> 17:27.360 They might, they're going to be on the hook for the liability in the user. 17:27.360 --> 17:37.400 Now it turns out that shortly thereafter, Congress passed another law, the Digital Millennium Copyright Act, which does give platforms some protection. 17:37.400 --> 17:51.140 But the reason I've put up this slide is this is a tiny piece of a tiny section of just one of pieces of the Digital Millennium Copyright, Section 512 alone, and this isn't particularly readable. 17:51.200 --> 17:53.600 It's not even particularly understandable. 17:53.600 --> 17:57.400 And companies go bankrupt trying to litigate. 17:57.400 --> 18:10.440 There is a terrible situation like if you think that it's a problem that YouTube is so big, you should know that there used to be a competitor to YouTube that got sued for copyright infringement allegedly in their users content. 18:10.440 --> 18:17.180 And they went bankrupt in the process of litigating it to the point where the Ninth Circuit Court of Appeals actually said there's no infringement. 18:17.180 --> 18:18.300 They're not liable for this. 18:18.300 --> 18:19.620 They should have been just fine. 18:19.620 --> 18:27.600 At that point, the company was gone and we lost a competitor, which now we're missing because it actually might be nice to have more competition online. 18:27.760 --> 18:34.980 So when we think about messing with these liabilities or making these liabilities complicated or conditional, this is what happens. 18:34.980 --> 18:38.240 We tend to obliterate them, and that's not what we want to have happen. 18:39.320 --> 18:46.440 There was also an exception from the very beginning that said there is no effect on federal criminal law. 18:46.440 --> 18:54.820 So that means that if the thing that was wrong with the content is that it violates a federal criminal law, also section 230 wouldn't apply. 18:54.940 --> 18:56.540 And for a while that was fine. 18:56.540 --> 18:59.140 I basically sort of said like child porn. 18:59.660 --> 19:00.620 That's bad. 19:00.620 --> 19:03.220 That tends to violate federal criminal law. 19:03.760 --> 19:09.580 And so we would impose on the platforms that they needed to do something about it. 19:09.680 --> 19:12.560 But that was one ask of a platform. 19:12.560 --> 19:29.540 And it's difficult and it's got some issues tied up in how do you do that effectively, but basically it was a single ask, our most important ask of what we could ask for a platform, and at least they could focus their resources on that and not have to be dragged into any other things that they might have to look for. 19:30.300 --> 19:34.680 But then things changed recently. 19:35.000 --> 19:42.780 A couple years ago, Congress put forth the first significant change to section 230 since it was passed. 19:42.940 --> 19:48.720 You might have heard it called SESTA, but by the time it got passed, it was referred to as FOSTA. 19:48.920 --> 19:53.100 And FOSTA basically said, you know, there's another type of bad content. 19:53.100 --> 19:55.400 We don't want sex trafficking either. 19:55.400 --> 19:56.060 And that's true. 19:56.060 --> 19:57.240 That's bad. 19:57.960 --> 20:01.720 But they poked a big hole in section 230. 20:01.720 --> 20:07.500 And one of the things to note, just looking at the slide, is they poked a very complicated hole in section 230. 20:08.900 --> 20:11.160 That's hard to read and understand. 20:11.160 --> 20:15.660 And you can see those hyperlinks are pointing you to other areas of the U.S. 20:15.660 --> 20:18.040 Code where they've put another criteria. 20:18.520 --> 20:27.940 It's made platforms really unsure about what sort of content is okay for them to facilitate and what sort of content is not. 20:27.940 --> 20:36.120 And as a result, they've started taking down or making themselves unavailable for content that would be totally lawful. 20:36.240 --> 20:43.200 A very famous example is Craigslist took down entire sections of classified ads, like dating ads. 20:43.200 --> 20:46.360 Dating ads are... they've been around for ages. 20:46.540 --> 20:47.880 Couples have met that way. 20:47.880 --> 20:53.360 There's nothing inherently wrong about an ad looking for dates and matchmaking. 20:53.440 --> 21:04.960 But because this is sort of adult content that can, you know, tip into the period, Craigslist quite reasonably got scared because they didn't really know, I think we might be held responsible for it. 21:04.960 --> 21:07.280 And that's just too big a chance to take. 21:07.540 --> 21:17.160 The other thing that this law did is it made it somewhat unsafe, and actually more than somewhat, it made an unsafe for a platform to take stuff down. 21:18.000 --> 21:22.340 Because... and we'll talk about this as we talk about other attempts to reform Section 230. 21:22.340 --> 21:27.420 They tend to mess with the idea of, like, if you knew it was there, then you're on the hook for it. 21:27.420 --> 21:30.100 So if you didn't know it was there, you might be okay. 21:30.100 --> 21:34.580 But then why should a platform take the time and energy to look for what might be there? 21:34.580 --> 21:37.180 Because it's just tempting trouble for themselves. 21:37.180 --> 21:42.500 So all of a sudden, now the platform is not a partner in dealing with the internet hygiene. 21:42.500 --> 21:43.920 And that's not constructive. 21:43.920 --> 21:47.580 It's not going to help us get the most good stuff or the least bad stuff. 21:48.080 --> 21:55.580 So remember these thoughts, how cripplingly expensive it is to be sued, even if the lawsuit is without merit. 21:55.840 --> 21:58.860 Remember that we want the most good stuff and the least bad stuff. 21:58.860 --> 22:03.100 And that carrots are more effective than sticks to helping us get that. 22:03.100 --> 22:11.080 And when we think about what some of these other laws are, you know, tested against that notion, and I think you'll find that most of these proposals fail. 22:11.080 --> 22:16.480 So with that, we're about to bring up Rihanna to talk to her. 22:16.480 --> 22:22.420 And one of the things we'll tip on is there is one proposal floating out there, which is okay. 22:22.580 --> 22:27.700 The first one, the examination of secondary effects for sex workers study. 22:27.780 --> 22:32.840 One of the things is that FOSTA kind of got rammed through where people said, I think we have a problem. 22:32.840 --> 22:45.860 But instead of really understanding the dynamic of that problem, they went ahead and changed the law around it without really understanding if that was a good solution, an effective solution, or one that didn't actually make the problem worse. 22:45.860 --> 22:52.220 Because it turns out there's a lot of evidence that it's actually made the problem they were trying to fix worse, worse. 22:52.500 --> 22:56.120 Women, there's a body count attached to FOSTA, sad to say. 22:56.120 --> 23:01.460 And this is a proposal to basically let's figure out what the effects are before we do anything else to it. 23:01.460 --> 23:02.760 But the rest is not okay. 23:02.760 --> 23:06.980 We're hearing executive orders trying to challenge Section 230. 23:07.140 --> 23:11.900 The DOJ has put out a guidance document that doesn't really understand it at all. 23:11.900 --> 23:19.400 The PACT Act is the most plausible where it seems to at least understand the statute, but it still creates problems. 23:19.640 --> 23:23.840 It seems like every week Senator Hawley is producing a bill to change it. 23:23.840 --> 23:30.740 And then there is the ERNED Act, which is sort of the bane of our existence, because it does more than change Section 230. 23:30.740 --> 23:33.000 I'm going to bring up Rhiannon now. 23:33.000 --> 23:36.060 We're going to close the sides and hopefully things don't break. 23:36.940 --> 23:40.260 And did I succeed? 23:41.460 --> 23:42.500 Okay. 23:43.300 --> 23:47.160 So, do we have Rhiannon? 23:47.180 --> 23:47.980 Yay! 23:47.980 --> 23:50.000 Okay, we're back to a single person. 23:50.400 --> 23:51.140 Great. 23:51.580 --> 23:58.200 Well, I was going to suggest we've gotten a few questions coming in over the Discord. 23:58.200 --> 23:59.440 Thanks, everybody. 24:00.340 --> 24:06.000 Maybe we could try knocking out a few of these, Kathy, before we get to talking a bit more about ERNED. 24:06.000 --> 24:07.180 Okay, excellent. 24:07.180 --> 24:07.760 Let's see. 24:07.760 --> 24:19.000 I see that... I'm seeing questions about moderation by bot and whether that is somehow bad or not, or that 230 speaks to it. 24:19.000 --> 24:22.560 And basically, 230 has a deliberately very light touch. 24:22.560 --> 24:30.320 It can apply to a whole bunch of systems in a whole bunch of contexts, and it's not very specific, which makes it very flexible. 24:30.320 --> 24:36.100 It was written in the mid-90s, and the internet has obviously changed quite a bit ever since, and it's flexible. 24:36.100 --> 24:49.440 It doesn't directly speak to AI, but if we stop and think about what AI is, AI, in theory, is the computer outsourcing of the brain of the people in the company. 24:50.200 --> 24:55.500 They have made decisions about how they want things moderated, and they deploy their bots to do it. 24:55.940 --> 25:02.440 The fact that you have technology pressing the button instead of a human being pressing the button, it doesn't change it. 25:02.440 --> 25:17.300 The core thing about Section 230 is that the intermediary is not responsible for the wrongfulness of the content that somebody else created, and the bots don't change that. 25:17.300 --> 25:24.260 They're still connected to the editorial discretion of the platform, and no, 230 is fine. 25:24.260 --> 25:26.240 It's not like it's an end run around it. 25:26.240 --> 25:29.780 They had an essential equation of who created the content and who made it wrongful. 25:29.780 --> 25:34.260 That's still core Section 230 stuff the way it was in the 1990s. 25:35.980 --> 25:40.800 And then I see the question about who gets to determine bad stuff. 25:41.340 --> 25:47.800 You know, I like to use the general term just to help people get their heads around it, but basically, the platform does. 25:47.800 --> 25:53.400 Because remember, I brought up that example of, like, you have a Facebook post and somebody posts a comment on it. 25:53.400 --> 25:58.860 You know, if they come and they insult your kid, you want to be able to delete that comment. 25:59.140 --> 26:03.780 Whereas if they come and they say your kid is awesome, you want to be able to leave that comment up. 26:03.780 --> 26:07.300 That is sort of the editorial discretion that the First Amendment protects. 26:07.960 --> 26:09.460 Same thing if you extend that. 26:09.460 --> 26:13.840 Instead of your own personal Facebook post, it applies to Facebook as a whole. 26:13.840 --> 26:24.360 They get to decide, because essentially, it offends the First Amendment to have an external government entity force anybody to carry speech. 26:24.700 --> 26:31.400 Minor exceptions to this, if we want to quibble, but they are very minor and they tend to be applied to things that are not the internet. 26:31.700 --> 26:37.680 So no, that's something that does get talked about a lot, where people are like, wait a minute, who gets to decide this? 26:37.920 --> 26:47.160 A private platform is a private platform, whether they're a large corporation of a lot of people who got together or a single person with their own post, it's the same equation. 26:48.040 --> 26:57.500 And the name of the YouTube competitor who got sued out of existence is, I don't know, the lawsuit ended up changed to shell, it was Veo Networks. 26:57.500 --> 26:59.920 That was the name of the video system. 26:59.920 --> 27:03.880 Veo Networks ceased to be, it is no longer a platform. 27:04.440 --> 27:09.920 So now my question is to Rihanna, what is wrong with the EARN IT Act? 27:12.660 --> 27:13.540 Thanks, Kathy. 27:13.540 --> 27:15.460 So everything is wrong with the EARN IT Act. 27:15.460 --> 27:16.980 Okay, end of stream. 27:16.980 --> 27:17.600 We're done here. 27:17.600 --> 27:18.420 We can all go home now. 27:18.420 --> 27:19.040 Thanks, everybody. 27:19.040 --> 27:19.820 Good night. 27:20.020 --> 27:35.380 No, so to be serious, the EARN IT Act is, I think, one of the most serious challenges and threats to online privacy, security, and freedom of speech that we have seen come across the transom in Congress for quite some time. 27:35.380 --> 27:44.240 It is one of, as Kathy mentioned, a bumper crop of anti-230 bills that would aim to narrow or even repeal Section 230 entirely. 27:44.240 --> 27:49.500 And sadly, it's also not the only anti-encryption bill out there right now, as I'll discuss in a moment. 27:49.840 --> 27:57.060 Kathy touched upon SESTA-FOSTA, which created a carve out from the 230 immunity with regard to sex trafficking offenses. 27:57.480 --> 28:06.000 In its current instantiation, the EARN IT Act bill is kind of similar, except instead of sex trafficking is child sex abuse material. 28:06.020 --> 28:08.560 Now, I'll call this CSAM for short. 28:08.940 --> 28:13.700 Child sex abuse material is a pernicious and difficult problem on the internet. 28:13.700 --> 28:24.420 It has been around ever since the days when people were allowed to transmit one file to another through some of the earliest services that you can think of, many of which I am too young to have used. 28:25.380 --> 28:33.640 And it's something that platforms have struggled with and that they've had to go through evolutions in their approaches to try and cut down on and prevent. 28:33.800 --> 28:41.440 The important thing to understand about CSAM is that it is illegal on the federal level, on the state level, across the entire planet. 28:41.440 --> 28:43.080 It's not legal anywhere. 28:43.500 --> 28:58.040 And as such, it is already subject to federal criminal statutes that impose duties on providers and platforms to report it and remove it when they learn about it on their services. 28:58.040 --> 29:03.660 That learning about it is important because it is what we would call an actual knowledge requirement in the law. 29:03.780 --> 29:08.100 Platforms are only required to take action about CSAM if they find out about it. 29:08.100 --> 29:09.860 That could be because a user reports it. 29:09.860 --> 29:26.520 It could be because they have an automated scanning service that looks at attachments that are in email or files that are uploaded to a cloud storage account to try and match against a database of hashes of known CSAM images. 29:27.100 --> 29:45.840 When they find something like this, they are required to report it to the National Center for Missing and Exploited Children, or NCMEC for short, which is a quasi-NGO, quasi, it's been deemed at this point to be an agent or an arm of the state, that handles reports of child, 29:45.840 --> 29:48.080 excuse me, of CSAM. 29:48.420 --> 29:54.220 Also, you might know NCMEC from the images on the side of milk cartons, the Have You Seen Me images. 29:54.220 --> 30:04.160 But I think at present, they're probably mostly associated with their really difficult role of having to process these reports that come in from providers. 30:04.160 --> 30:09.780 And providers do report CSAM that they learn about on their services millions of times a year. 30:10.580 --> 30:27.860 There's also federal requirements that say that the Department of Justice is supposed to provide funding for, conduct studies, conduct reporting with respect to all of these reports that are coming in with respect to trying to fight and minimize the abuse of children and other crimes against children online. 30:27.860 --> 30:29.240 They've been falling down on the job. 30:29.240 --> 30:37.420 They have not been filing their federally mandated reports most of the years that they have been required to do it over the last decade or so. 30:38.100 --> 30:53.820 And I think one of the motivations for the EARN IT Act bill is to try and point a finger away from DOJ's own shortcomings and away from how overwhelmed NCMEC is because they do not have the resources, they do not have the personnel they need in order to take on this task and deploy that platforms instead. 30:53.940 --> 31:01.180 And the way that they've chosen to do that is by taking kind of the, you know, the whipping boy do short, which is Section 230. 31:01.720 --> 31:07.480 Right now there's a lot of kind of popular discontent, I think, with Section 230, much of it ill understood. 31:07.480 --> 31:15.860 And so thank you to Kathy for trying to explain a bit better why Section 230 is a important force online. 31:16.580 --> 31:30.920 But by going after Section 230, it enables, I think, the DOJ to try and both direct attention away from its own shortcomings and also to promote an agenda that has long had for greater surveillance of online speech, and in particular to go after encryption, 31:30.920 --> 31:35.700 which is an area that I focus on in particular in my role at Stanford. 31:36.140 --> 31:46.260 So by curtailing immunity for under Section 230 for CSAM, there's nothing about that that is not something that DOJ can already do. 31:46.480 --> 31:53.880 Kathy discussed how one of the limitations on 230 immunity is that it already does not bar enforcement of federal criminal law. 31:53.880 --> 32:01.480 You even saw where she noted that child sexual abuse materials explicitly called out in that particular clause in 230. 32:01.480 --> 32:15.240 So it is already the case that if platforms are not complying with those aforementioned reporting duties when they learn about CSAM on their services that I mentioned, nothing at all is stopping the Department of Justice from penalizing those platforms. 32:15.240 --> 32:36.060 All Section 230 currently does is to bar state AGs, state attorneys general, from bringing criminal charges against platforms because Section 230 bars state criminal law, and it bars private plaintiff lawsuits by individual people with private right of action because those two are barred by the Section 230 immunity. 32:36.060 --> 32:49.960 So we're in a situation where providers are subject to statutorily mandated duties with respect to CSAM by all of evidence they are complying with those things, but there's a sense that they should be doing more. 32:49.980 --> 33:00.200 The fact that they are reporting millions of pieces of CSAM a year to NCMEC is taken not as evidence that they are finding and reporting things, but that they aren't doing enough, which is a bit curious. 33:01.240 --> 33:26.040 So the point of the Grant Act bill in its current instantiation since its introduction earlier this year kind of doesn't make sense under the name anymore because originally what the bill would have done would have been to make continued immunity under Section 230 contingent upon compliance by providers with some to-be-determined best practices. 33:26.040 --> 33:39.900 If you follow these best practices, you jump through the hoops, these would be set by an unelected commission of people headed by the Attorney General, then you could maybe continue to be eligible for continued Section 230 immunity for CSAM. 33:40.900 --> 33:43.960 Otherwise, you might just be sort of thrown to the wolves. 33:43.960 --> 33:51.980 That link between following the best practices set by the commission and continued immunity is taken out in the current version of ERNIT. 33:51.980 --> 33:56.740 Current ERNIT is much more like SESTA-FOSTA in that it abrogates the immunity full stop. 33:56.740 --> 34:06.380 It still has this commission, this 19-member commission, who are still charged with coming up with best practices for fighting CSAM online and protecting child safety online. 34:06.380 --> 34:16.500 However, those are now just non-binding voluntary recommendations that are not tied to any particular carrot or any particular stick under the bill. 34:16.500 --> 34:19.300 So there's nothing left to earn. 34:19.300 --> 34:21.240 There's no more immunity left to earn. 34:21.240 --> 34:41.340 And the ultimate upshot of all of this is because the removal of the 230 immunity bar opens up potential liability under what could be 50 or more state and territorial laws that deal with CSAM, there's a lot less certainty for platforms about what their duties are. 34:41.340 --> 34:46.420 Right now, they know what they have to do as specified under the federal reporting law. 34:46.420 --> 35:01.360 But under the ERNIT Act, there could be a variety of state laws that might open them up to liability for CSAM and to private civil lawsuits and to state AGs going after those platforms under what could be much lower bars. 35:01.360 --> 35:12.720 Going back to what I mentioned about the actual knowledge requirements under the federal reporting statute, that's a fairly permissive bar for platforms. 35:12.720 --> 35:14.720 It's not that low. 35:14.720 --> 35:33.540 Whereas, we know from a survey that a group called NetChoice has done of state CSAM laws, that some state level CSAM laws impose liability on those who trade or host or transmit CSAM under a much lower standard, such as negligence or recklessness. 35:33.580 --> 35:46.360 So platforms could easily potentially be held liable for CSAM on their services if they negligently or recklessly host it, transmit it, allow it to be promoted or presented or distributed, etc., etc. 35:47.800 --> 35:52.480 You might be asking yourself, wait a second, doesn't that just mean the ability to transmit files? 35:52.480 --> 35:53.700 Why, yes, it does. 35:53.760 --> 36:07.620 All of the things that we do online, such as talking to each other, such as sending files to each other, this is what the internet is for, those can also be used in ways that can lead to the exploitation of children or that can involve the transmission of CSAM online. 36:07.620 --> 36:24.660 So by potentially exposing platforms to liability for negligently or recklessly transmitting CSAM, they're really potentially exposing platforms to liability for hosting user contributing content at all, for the ability to transmit files, for the ability to speak online and speak to other people, 36:24.660 --> 36:28.460 some of whom may or may not be over the age of 18. 36:28.460 --> 36:33.240 So the chilling effect of this bill's potential cannot be overstated. 36:33.240 --> 36:41.240 We already saw with regard to FOSTA how it led to the shutdown of large swaths of completely legal speech on the internet. 36:41.240 --> 36:45.260 You may recall that Craigslist immediately shut down its entire personal section. 36:45.260 --> 36:49.860 They also shut down their section that allowed for therapeutic services. 36:49.860 --> 37:07.760 So massage therapists were suddenly out of that part of their livelihood that came from finding clients on Craigslist, for fear that somewhere in those personals ads, or somewhere in their massage therapy ads, there might be somebody who was being trafficked and exposed those platforms Craigslist to liability under SESTA-FOSTA. 37:07.760 --> 37:16.220 Here, similarly, we can predict that there would be a similarly large chilling effect on the ability of people to speak to each other online. 37:16.260 --> 37:20.580 Because again, what if any particular file might sneak through? 37:20.580 --> 37:29.340 What if some communication between two users on the service might be exploitative or enticement between an adult and a minor? 37:29.340 --> 37:38.500 These are all things that platforms are going to be wrestling with in the event the EARN IT Act passes, because they now may be exposed to liability under 50 or more state laws. 37:39.360 --> 37:46.340 So the fact that there have been some changes that were made to the EARN IT Act don't really fix the core problems with the bill. 37:47.120 --> 37:58.620 When we think about how this bill relates in particular to encryption, the changes that were made between the early version of the bill and the current version of the bill were intended to placate concerns somewhat. 37:58.620 --> 38:01.500 I'm here to tell you, it doesn't fix it. 38:01.500 --> 38:17.140 There's an amendment offered by Senator Leahy that was intended to protect encryption and take away, basically restore immunity from suit for the use of encryption, for deploying it in encryption, for providing device encryption, and so forth. 38:18.820 --> 38:22.000 It's not really enough, though, in my opinion. 38:22.000 --> 38:38.260 Both because it does not protect other forms of providing user security and would still allow lawsuits to arise if somebody who was determined to bring a criminal charge as a state AG or a private plaintiff were determined to bring a lawsuit. 38:38.260 --> 38:45.800 All they'd have to do would be to come up with some reason for suing, let's say WhatsApp, let's say Signal, that wasn't directly about the encryption . 38:45.800 --> 39:02.060 Any other pretext that they could find or any other security measure that was in use, even, say, data minimization practices, for example, if that would potentially get in the way of fighting CSAM, maybe that's something that you use as the basis for trying to make a platform liable. 39:02.740 --> 39:19.640 And on the other hand, I don't think that the way the Leahy amendment that is supposed to protect encryption is structured would necessarily mean that platforms would be safe from being pressured or induced into introducing other forms of user surveillance that don't touch the encryption. 39:19.640 --> 39:34.560 One of the moves that we have seen in the encryption debate over the last few years has been, as we've pushed back against proposals to weaken encryption, to insert mandatory backdoors into encryption, we have instead seen law enforcement proposals come out that say, 39:34.560 --> 39:35.940 well, how about this other idea? 39:35.940 --> 39:38.260 It doesn't involve breaking the encryption at all. 39:38.260 --> 39:58.560 And those include things like the so-called ghost user proposal proposed by GCHQ, the spy agency in the UK, which would silently add a law enforcement user to an otherwise end-to-end encrypted conversation while suppressing notifications to the people who are present in the conversation that this silent ghost law enforcement user has been added. 39:58.560 --> 40:00.260 It doesn't touch the encryption, right? 40:00.260 --> 40:06.840 It's still a way of undermining the privacy guarantees and security guarantees that you expect from using an end-to-end encrypted app. 40:06.840 --> 40:21.460 That's just an example of how the Leahy amendment doesn't really do enough to protect encryption and to keep providers from being disincentivized to continue offering end-to-end encryption and other security measures to their users. 40:22.220 --> 40:25.260 So I know I get a little head up about these things. 40:25.600 --> 40:27.280 It's hard to restrain. 40:27.280 --> 40:28.520 I could talk about this all day. 40:28.520 --> 40:42.660 But before I wrap up and we go on to maybe having a bit more of an actual fireside back and forth, I want to mention, I noted that Section 230 bills aren't the only threat to our encryption and privacy and online speech rights that are out there. 40:42.660 --> 40:48.220 There's another companion bill that was introduced by the same sponsor, Senator Graham, of the EARN IT Act. 40:48.220 --> 40:54.560 He introduced it just a few days before releasing the amended version of the EARN IT Act a few weeks ago. 40:54.560 --> 41:01.760 It's called the Lawful Access to Encrypted Data Act, or I like to say the L is silent, and it's the Awful Access to Encrypted Data Act. 41:01.760 --> 41:04.460 And that is just a flat-out backdoor bill. 41:04.460 --> 41:09.020 It is something that we have been waiting for, hoping wouldn't come out. 41:09.140 --> 41:10.860 Dreading, I think, is the word. 41:10.860 --> 41:12.820 Dreading is absolutely the word. 41:12.840 --> 41:31.440 It is really the full frontal nuclear assault on your ability to just be able to use encryption and move on with your life that we have gotten used to in recent years as encryption by default has become the norm in the devices that we use and in the apps and services that we use. 41:31.440 --> 41:54.260 So what is so dangerous about the LAED Act, or the Laid Act if you want to sound a little bit naughty about it, is that it would require basically any sort of larger provider, anybody who sold a million devices in the United States in a given year, or who has a million monthly active users in any given month dating back to 2016, 41:54.260 --> 42:07.520 so even retroactively, would have to proactively re-engineer their services to be able to comply with warrants or other legal process they received from law enforcement to decrypt information that was encrypted. 42:07.520 --> 42:10.840 They'd have to proactively redesign to build a backdoor in. 42:10.840 --> 42:33.720 If you are a smaller provider that's not part of that million plus club, let's say a very small handset maker or the provider of an end-to-end encrypted app or other encrypted service that does not have a million users, you still might, you're still covered by this bill because then the Attorney General would be able to serve a notice on you telling you to re-engineer your service and add that backdoor into your service, 42:33.720 --> 42:37.900 even though you would not be forced to re-engineer it proactively. 42:37.900 --> 42:56.640 So this bill was introduced, as I said, same sponsor, right before the EARNET amendments were made, clearly intended to make EARNET look more reasonable by comparison and make it look less like an attack on encryption by comparison because this is an overt rather than covert attack on encryption. 42:56.640 --> 42:58.820 But the lesser of two evils is still evil. 42:58.820 --> 43:02.700 Like the EARNET Act is bad and the LAED Act is worse. 43:02.700 --> 43:06.240 It's not as though Congress members have to pick between one or the other. 43:06.240 --> 43:16.980 They could just pick neither and, oh, I don't know, focus on making sure that millions of Americans don't become fucking homeless, you know, in the next few weeks and, you know, do something to fix the epidemic that is ravaging this country. 43:16.980 --> 43:24.280 Instead, lawmakers have been deciding to focus on what really matters, which is Google and Facebook are apparently annoying to some people. 43:24.280 --> 43:25.100 Whatever. 43:25.160 --> 43:28.860 Go and focus on what really matters, Congress, and start leaving the internet alone. 43:28.860 --> 43:30.580 Leave my right to encrypt alone. 43:30.580 --> 43:33.040 So please contact your Congress members. 43:33.040 --> 43:34.540 I wrote it down here. 43:34.540 --> 43:40.000 I'm making this very small so that nobody can just, you know, photoshop me and say something stupid. 43:40.100 --> 43:41.500 Know EARNETact.org. 43:41.500 --> 43:42.600 That's where you need to go. 43:42.600 --> 43:43.840 Call your Congress members. 43:43.840 --> 43:50.240 Call your senators in particular, because the EARNET Act bill has advanced fairly far within the Senate already. 43:50.240 --> 43:52.860 It's already gotten out of the committee in which it was introduced. 43:52.860 --> 43:55.580 It passed out unanimously out of that committee. 43:55.960 --> 44:12.560 And there has been efforts just this week by Senator Graham, its lead sponsor, to try and force it through by what's called unanimous consent so that it would pass out of the Senate without all of those members of Congress actually having to stand up there on the floor of the Senate and be seen on C-SPAN voting in favor of this terrible, 44:12.560 --> 44:13.700 destructive bill. 44:13.700 --> 44:15.820 So please contact your senators. 44:15.820 --> 44:17.340 Contact your House reps, too. 44:17.340 --> 44:22.340 And make sure that while you're at it, you tell them to oppose the Lawful Access to Encrypted Data Act as well. 44:22.340 --> 44:24.060 That bill hasn't gotten as far in the Senate. 44:24.060 --> 44:25.400 It hasn't gotten out of committee yet. 44:25.400 --> 44:35.220 But it has already had a companion bill introduced in the House by one of the same people who was behind the SESTA-FOSTA precursors and is also a threat to online speech. 44:35.220 --> 44:40.040 So make sure, tell your senators, tell your House reps, oppose both of these bills. 44:40.620 --> 44:46.740 And now that I can just stop being a little verklempt, I think we can move on and have this chat. 44:46.740 --> 44:48.560 We're not going to completely move on. 44:48.560 --> 44:52.280 We're going to tease out some of the reasons why this is bad. 44:52.280 --> 44:59.600 Because I think we need to, the same way we need to understand how 230 works, I think we need to understand what went wrong and what goes wrong here. 44:59.740 --> 45:04.640 So I think one of the things that's come up is, this is not a small thing. 45:05.060 --> 45:11.440 That basic thing we talked about, about what should Congress want from the internet providers? 45:11.440 --> 45:14.880 Them to help us get the most good stuff and the least bad stuff. 45:14.880 --> 45:19.160 And one of the things you talked about is that it does the exact opposite. 45:19.160 --> 45:36.080 You talked about how it will make platforms completely afraid to remain available to intermediate good content because anything could set off the liability threat, especially now when we're adding all those different states. 45:36.080 --> 45:52.440 So remember that one of the things I pointed out was there was a preemption clause in section 230 because it would be a mess if you had 50 different state attorney generals with 50 different opinions about how things should be, essentially getting to decide what internet policy was. 45:52.440 --> 46:09.120 And so you point out that the internet essentially does away with that preemption and now puts an awful lot of power among individual state attorney generals to essentially censor the internet and it'll essentially become a situation of the least common denominator of, 46:09.120 --> 46:14.340 you know, that may be great that 49 states are like, yeah, yeah, whatever, be reasonable, it's fine. 46:14.340 --> 46:23.260 But if one state is like, I'm coming to get you if you do something that offends what I've declared to be the policy so that I can get re-elected next term, that's going to be a problem. 46:23.260 --> 46:41.260 And basically the rest of the United States, their internet experience will get shaped by that single, it's the least common denominator thing that if one state, you know, wants to have something that's irrational, but it's on the books, they've got the power now to do it because the section 230 protection goes away. 46:41.260 --> 46:49.180 And when that protection goes away, the choices that platforms make to have the most good stuff and to minimize the bad stuff goes out the window. 46:49.480 --> 47:06.880 You also talked about how it messes up them being partners and getting rid of the bad stuff, because right now we've already left because of that carve out for federal criminal law, we've told them, go ahead and, you know, we want you to get rid of the child porn, 47:06.880 --> 47:08.280 get rid of that. 47:08.880 --> 47:18.760 Platforms are making an effort to do it, but now you've sort of created a weird incentive of like, if you look for it, now you might be in trouble for having found it. 47:18.760 --> 47:22.380 So why would a platform now want to look for it? 47:22.380 --> 47:24.820 That seems like it would just be begging for trouble. 47:24.920 --> 47:28.000 Why wouldn't they just encrypt everything so that they could see nothing? 47:28.240 --> 47:30.140 And now we mess up your encryption. 47:30.240 --> 47:50.400 And yeah, so I mean, let's talk a little bit about the politics because I think one of the things and one of the reasons why we wanted to do this talk is because section 230 has become this very hot topic that everybody is banding about because it's become the easy scapegoat for anything that's wrong online and anything that's wrong online that's affecting the rest of our world, 47:50.400 --> 47:53.920 which is difficult these days for a variety of reasons. 47:54.520 --> 48:03.020 So you have people who have differing agendas, but we're all kind of unhappy these days and we've now found something to blame. 48:03.020 --> 48:18.040 And so you end up with these weird bedfellows of, Ernact Act is sponsored in part by Senator Graham, who is a Republican, but it's also equally pushed by Senator Blumenthal, who is a Democrat. 48:18.040 --> 48:29.420 How on earth did we get a Republican and Democrat in this incredibly polarized world to agree that messing up section 230 with this law is a good idea? 48:29.420 --> 48:33.600 How does this have appeal to both parties at the same time? 48:33.600 --> 48:38.220 And does that maybe suggest for each camp that maybe they should rethink it? 48:38.220 --> 48:46.360 Because if the thing they want is making the other side happy, maybe that should cause some concern about what it is they actually want. 48:46.360 --> 48:50.240 I mean, the easy answer there is this is a bill that's supposed to be about child safety. 48:50.240 --> 49:03.080 It's not actually going to help protect children on the internet because if major platforms like Facebook or like Twitter now know that they are exposed to liability unless they get a lot more intrusive in terms of the amount of information that they collect from users, 49:03.080 --> 49:20.260 in terms of the surveillance they do of their users, then everybody who is treating in this material on these major platforms is just going to be forced off and they're going to be using Tor hidden services and they're going to be on other sites that don't comply with federal CSAM reporting laws because they exist in order to host this kind of material, 49:20.260 --> 49:26.980 don't qualify for section 230 because they are themselves engaged in violating the law, and don't care. 49:26.980 --> 49:31.120 So it's not going to achieve the stated goal of helping protect children. 49:31.120 --> 49:41.800 By the way, this does not provide the sorts of services for housing stability, for removing children from abusive homes, all the stuff that you could try to do to prevent abuse from happening in the first place. 49:41.980 --> 49:45.280 It's just meant to be another thing of the children bill. 49:45.280 --> 49:48.960 And so it doesn't achieve those goals that it sets out to do. 49:48.960 --> 50:01.620 But by being out there as a bill that's purporting to protect child safety, it makes it much more bicameral friendly in terms of getting a bipartisan buy-in from each side of the aisle. 50:01.680 --> 50:08.080 And it makes it a lot harder for members of Congress to vote against it, because who wants to be seen as voting against something that helps the children? 50:08.080 --> 50:09.060 Do you hate the children? 50:09.060 --> 50:10.260 You must hate the children. 50:10.360 --> 50:12.120 Well, it prompts the two questions. 50:12.180 --> 50:16.420 One, this isn't actually going to help them, it's going to hurt them, correct? 50:16.820 --> 50:17.540 Right. 50:17.680 --> 50:18.000 Yeah. 50:18.000 --> 50:18.280 Okay. 50:18.280 --> 50:23.640 So first of all, it's not going to accomplish what it set out to do, just like FOSTA didn't accomplish what it set out to do. 50:23.640 --> 50:28.700 But I think the big problem is, yeah, who wants to vote against keeping children safe? 50:28.700 --> 50:36.740 So if it's got that banner bumper sticker, think of the children, vote for this, it's a very difficult thing politically to say no to. 50:36.780 --> 50:42.700 But that means that it becomes a really convenient cover for people who have other agendas. 50:43.600 --> 50:48.200 Because they've now made the poison pill that the other side can't say no to. 50:48.240 --> 51:00.660 And I mean, that was happening even with SESTA-FOSTA, where people who had agendas for why they wanted to get Section 230 off the books, were able to push that. 51:00.660 --> 51:04.420 Do you have any thoughts about what some of these agendas are? 51:04.640 --> 51:06.080 I mean, I absolutely agree. 51:06.080 --> 51:15.820 And you know, there's an organization of sex workers, and allies who have been really vocally working against EARN IT ACT, as they did against SESTA-FOSTA called Hacking and Hustling. 51:15.820 --> 51:26.420 And one of the slogans that they use when they are describing both the impact of SESTA-FOSTA and the likely impact of EARN IT, is that it punishes people because we want to punish platforms. 51:26.420 --> 51:29.460 Not you and I, we, but you know, the general... 51:29.460 --> 51:31.100 We personally, no, we're cool with them. 51:31.100 --> 51:31.360 Yeah. 51:31.680 --> 51:34.480 Don't punish people because you want to punish platforms. 51:34.480 --> 51:41.680 If the idea is somehow that Google or Facebook or whoever is too big for their britches, or they're making too much money, our tax code allows them to make it. 51:41.680 --> 51:45.280 Or they're somehow acting above the law, the law permits them to do everything that they're doing. 51:45.280 --> 51:46.620 That's what Section 230 does. 51:46.720 --> 51:56.100 Then, you know, you pass a bill that ends up making it more dangerous for sex workers to be alive, to exist, to work. 51:56.100 --> 52:03.140 And that will make it harder, potentially, to prosecute CSAM traders if this bill passes. 52:03.200 --> 52:07.200 I agree with you that there is an ulterior agenda in each of these cases. 52:07.200 --> 52:24.660 And one thing that I thought was really problematic, in particular, about the use of the EARN IT bill, as what was clearly just a Trojan horse for trying to ban encryption, or at least strongly discourage it, was that we're talking about such a wide range of services that would be covered when really we're talking about a couple of different things. 52:24.660 --> 52:31.180 The thing that makes people angry about the way that the internet kind of sucks is the stuff that's not encrypted. 52:31.180 --> 52:38.740 It's the comments that you see on Twitter, it's the comments you see on YouTube, it's why we have emoji-only chat in our Twitch stream this weekend. 52:39.040 --> 52:47.440 You know, it's the things that are happening in unencrypted spaces, I think, are largely what motivate people being just kind of sick and tired of what the internet feels like these days. 52:47.440 --> 52:57.640 And yet, the thing that is being attacked by efforts to sort of be a backdoor to an encryption backdoor mandate is encrypted private communications. 52:57.640 --> 52:59.060 It's private spaces. 52:59.060 --> 53:09.100 So it's trying to take this popular anger at something that doesn't have to do with the private communications that we have in one-on-one spaces and sort of misdirect it. 53:09.100 --> 53:19.740 Again, this is all a big game of misdirection between the DOJ, which I'm convinced largely wrote both of the bills I'm talking about, especially the second one, to try and push that agenda. 53:19.740 --> 53:23.740 And, you know, I do believe overall that, you know, it would be very difficult. 53:23.740 --> 53:29.660 We talk in the encryption debate sometimes about finding some sort of middle ground proposal or some sort of compromise. 53:29.840 --> 53:39.580 But it seems like the only compromise, the only middle ground is only ever demanded from the people who have the right and now the ability to protect themselves and to exercise their rights. 53:39.580 --> 53:42.620 It's never a law enforcement offering to give ground. 53:42.620 --> 53:44.120 I'm very dubious. 53:44.720 --> 53:52.880 I mean, I've been called an absolutist, but the reason I'm an absolutist is I'm really dubious about that there's a way to have a compromise. 53:52.880 --> 54:00.480 I mean, I pulled up, I did this slide about the Digital Millennium Copyright Act, which in theory is a type of conditional safe harbor. 54:00.480 --> 54:05.140 If you do X, Y, and Z, okay, then you're not in trouble for your third party content. 54:05.140 --> 54:07.100 But it's a mess. 54:07.180 --> 54:09.420 It's very difficult to figure out the conditions. 54:09.420 --> 54:11.100 It's very difficult to meet them all. 54:11.100 --> 54:12.940 It's not necessarily realistic. 54:12.940 --> 54:32.880 The point I didn't make for that is how much censorship of lawful speech happens because we've made this, the platform protection so conditional where people send garbage takedown notices that the platforms really have no choice but to take down, even though the copyright claims behind them are garbage, 54:32.880 --> 54:35.460 sometimes because they're not even really copyright claims at all. 54:35.460 --> 54:43.340 But a platform handling voluminous content with and potentially expensive liability doesn't have the ability to say no to them. 54:43.340 --> 54:44.620 I mean, sometimes they do. 54:44.620 --> 54:47.720 But it's rare, because it can't happen all that often. 54:47.720 --> 54:49.300 It's just not practical. 54:50.120 --> 54:55.380 So I don't see how we can, this is kind of an all or nothing thing. 54:55.380 --> 54:59.780 And that's why I go back to the, do you want the most good stuff and least bad stuff? 54:59.780 --> 55:03.960 I think there's really only one way to achieve it. 55:03.960 --> 55:13.720 And that's by aligning the incentives of we have more tools in our regulatory tool belt to regulate, it doesn't have to be penalty based. 55:13.720 --> 55:15.880 And somehow I think we're not good at that. 55:15.880 --> 55:28.340 I think we don't understand how to regulate effectively by making parties partners, where acting consistent with their own interests happens to get them to do what we want. 55:28.860 --> 55:36.020 The reason the internet sprouted and did so well in 20 years is because that's how we regulated because that's how we made it work. 55:36.020 --> 55:39.880 And that's how we went from basically no internet to an awesome internet. 55:39.880 --> 55:47.780 I mean, it's got wrinkles, but you can connect the whole world, we can connect like an infinite amount of knowledge and ideas and make connections. 55:48.580 --> 55:51.580 By and large, it is incredibly awesome. 55:51.780 --> 56:03.280 Maybe part of the problem we have is that, you know, if you've got 7 billion people on, you know, connected to each other, some of them suck, some of them are not good people, some of them say terrible things. 56:03.280 --> 56:11.840 And when these terrible things happen, we become hyper aware of them, because a we now can be aware of them, because the internet is providing us with that information. 56:12.140 --> 56:17.860 And they end up with this outsized effect, like, oh, my gosh, I can't believe what that person did. 56:17.860 --> 56:19.080 And we're upset about it. 56:19.080 --> 56:23.620 And we look at what existed that allowed the bad thing to happen. 56:23.620 --> 56:24.840 Well, that was the internet. 56:24.840 --> 56:27.360 And the internet got to exist because of Section 230. 56:27.360 --> 56:30.480 So therefore, 230 is bad, let's get rid of it. 56:30.480 --> 56:37.020 And it's really hard to sort of call time out to reasonable people and say, hang on, maybe that's not the best way to think about it. 56:37.020 --> 56:46.260 Although, that's why we're doing presentations like this, to kind of walk people through the thinking, so that they understand that, like, no, this is really important and really valuable. 56:46.260 --> 56:52.040 And maybe the things that are going wrong are things that we should target our regulation to. 56:52.040 --> 57:07.280 That if we're upset about something, let's carefully define the problem and come up with an appropriate solution that helps, you know, fix the thing that's wrong without just taking away the whole thing that's created the internet, good things and bad things, 57:07.280 --> 57:09.320 because now we won't have the internet at all. 57:09.320 --> 57:09.640 Right. 57:09.640 --> 57:13.060 Human terribleness has been around since well before there was an internet. 57:13.060 --> 57:16.280 People have been terrible for as long as they have been people. 57:16.660 --> 57:26.820 Nevertheless, I like your point about how we tend to sort of focus on the bad parts without thinking about the vast majority of perfectly fine content. 57:26.820 --> 57:28.400 The same is true of encryption. 57:28.400 --> 57:47.620 You know, I feel that the focus on the use of encryption to to commit crimes or to hide evidence of crimes gets blown all out of proportion compared to how rare the instance of crime really is when you compare it to the fact that more than one out of seven humans on this planet use WhatsApp. 57:47.620 --> 57:50.200 One billion people aren't criminals. 57:50.200 --> 57:50.300 Right. 57:50.300 --> 58:01.060 So most of the usages, most of the people using encrypted apps, most of the people using encrypted devices, which at this point is basically everybody, just by default, most of that is not for crime. 58:01.060 --> 58:16.160 And so we lose sight of placing these things into context and thinking in more big picture about what the cost benefit analysis is of trying to increase surveillance and reduce free speech rights for everybody through bills like this. 58:16.260 --> 58:19.680 I want to make sure we get to some of the questions that have been coming in. 58:19.680 --> 58:22.320 We have maybe about five or ten minutes left to take questions. 58:22.320 --> 58:25.060 I know we've run over a little bit, but it's the end of the day. 58:25.060 --> 58:26.400 There's no rules. 58:27.360 --> 58:29.380 Well, wait till Congress gets at it. 58:29.380 --> 58:35.640 But OK, so I think some of these questions now, some of them are sort of technical to 30 questions. 58:35.960 --> 58:43.080 I almost can globally say, yes, 230 totally applies, although, yeah, if we can get to them, we'll be more specific. 58:44.480 --> 58:48.180 Actually, the most interesting one is one that I've been putting forth. 58:48.180 --> 58:51.820 And since I'm a speaker, I can say that somebody asked a question about 230 and Amazon. 58:52.000 --> 59:00.540 One of the things that's starting to happen is some of Internet services have started to look more like actions than speech, because people sell things on Amazon. 59:00.540 --> 59:02.880 And is that a different sort of activity? 59:02.880 --> 59:13.300 And I've argued in briefs before courts that basically it still boils down to speech, that if you're on Airbnb, what you're really saying is, I have an apartment to rent. 59:13.340 --> 59:19.700 That may be the only thing you're really saying on Airbnb, because it's a very hyper-focused platform. 59:19.800 --> 59:26.220 But again, that whole slide about 230 is kind of general in its application, and it applies to a whole bunch of things. 59:26.220 --> 59:33.820 It can apply to Amazon when you regard that a third-party seller is really a seller saying, I have a thing to sell. 59:33.820 --> 59:49.760 And that when we start imposing liability for attacks to the sale, that ripples through and affects every platform that hosts every type of expression, because there's no real contingency of whether 230 applies to a broad-based platform that takes all sorts of speech, 59:49.760 --> 59:55.360 or a narrow-focused platform that takes very specific, I'm only selling this kind of thing type of speech. 59:55.500 --> 59:58.360 So that's a lot of the 230 things. 59:58.660 --> 01:00:05.980 On the EARNET side, people are asking, and maybe this is the way to phrase the question, if EARNET passes, what happens? 01:00:06.180 --> 01:00:07.940 Does signal go offshore? 01:00:07.940 --> 01:00:09.140 Does something go offshore? 01:00:09.140 --> 01:00:10.560 Does the internet die? 01:00:10.560 --> 01:00:12.140 I mean, it kind of died. 01:00:12.140 --> 01:00:20.000 Well, it didn't completely die with FOSTA, except it really took a very severe hit, and so did an awful lot of the users who used it. 01:00:20.000 --> 01:00:24.420 Like, that body count of sex workers is significant and real. 01:00:24.780 --> 01:00:27.700 So let me ask you, what happens if EARNET passes? 01:00:27.900 --> 01:00:38.840 If EARNET passes, I think we will immediately see even more shutdowns of portions of services, or even maybe entire services, just as we saw immediately the moment SESTA-FOSTA was signed into law. 01:00:38.840 --> 01:00:46.880 I think we'll see a lot of litigation, including over this encryption safe harbor that Senator Leahy tried to insert in. 01:00:47.360 --> 01:01:00.860 It basically creates kind of a license to litigate, where because it is narrowing what should be a broad immunity, and which already takes quite a lot of litigation often to just get to the point of saying, no, Section 230 bars this lawsuit and kicking it out of court. 01:01:00.860 --> 01:01:12.560 I think we'll see a lot of litigation over whether a lawsuit is trying to hold a platform accountable for CSAM on its service on the basis of encryption or some other thing. 01:01:12.560 --> 01:01:14.480 If it's encryption, then the lawsuit is barred. 01:01:14.480 --> 01:01:19.720 If it's some other pretextual reason that the plaintiff or the state AG can come up with, then it's not going to be barred. 01:01:19.720 --> 01:01:30.440 So we'll see platforms drawn into quite a lot of litigation, and that's for the ones that dare to continue providing any kind of functionality for user-generated content at all. 01:01:30.440 --> 01:01:33.420 I think we may see small platforms start to shut down. 01:01:33.520 --> 01:01:39.160 It's true that Signal has threatened to go offshore and move to another jurisdiction. 01:01:39.160 --> 01:01:41.400 This is one of the limitations with this bill. 01:01:41.400 --> 01:01:46.340 I don't know how it could possibly purport to touch open source tools. 01:01:46.340 --> 01:01:51.160 Most of the encrypted tools out there are either open source or housed overseas outside of U.S. 01:01:51.160 --> 01:01:52.040 jurisdiction. 01:01:52.040 --> 01:01:58.980 So these tools would still be available, and those who want to use encrypted tools for bad acts would flock to using them. 01:01:58.980 --> 01:02:01.600 There's another question in the stream about that. 01:02:02.300 --> 01:02:16.380 But the people who are just everyday average users probably won't take the time to seek out services that have not been intentionally weakened in order to try and avoid the specter of liability that EARNET would raise, in the event that that is in fact what platforms do. 01:02:16.380 --> 01:02:35.980 I think we'll see a difference between the response, as you were talking about earlier, between the large well-resourced platforms that can handle the deep pockets needed to face this kind of litigation and try to make that calculation that they would rather protect their users and continue keeping these kinds of security protections in place and just deal with, 01:02:35.980 --> 01:02:39.200 you know, paying for the stuff that they don't catch. 01:02:39.260 --> 01:02:44.860 And we might see smaller services either shutting down or, as mentioned, potentially offshoring. 01:02:44.860 --> 01:02:45.940 So we'll see. 01:02:45.940 --> 01:02:55.080 But all the uncertainty that that opens up is just another reason to, like, keep fighting against this bill, because everything about the world is uncertain right now. 01:02:55.080 --> 01:03:00.380 I would like to know that, like, my ability to just use my phone and use whatever you signal is still certain. 01:03:00.380 --> 01:03:07.540 So I think this boils down to that, you know, understand how the law works and the basic mechanics of it. 01:03:07.540 --> 01:03:09.400 That's how 230 works. 01:03:09.400 --> 01:03:12.900 And then definitely understand how the statute messes with that. 01:03:13.160 --> 01:03:19.660 I think it boils down to also, like, I asked the question about, like, what is the political motivation behind some of these things? 01:03:19.660 --> 01:03:29.420 And to the extent that there's genuine displeasure with some of the larger internet companies, the point you're making also is that they'll probably be able to survive. 01:03:29.420 --> 01:03:33.520 But if one of the things we really want to have, like, if we don't like Google, we really need a new Google. 01:03:33.520 --> 01:03:35.360 And this is not going to help us get a new Google. 01:03:35.360 --> 01:03:40.980 This is going to help us double down and entrench the Google with everything we don't like about it doubled down. 01:03:41.020 --> 01:03:43.700 So it doesn't actually solve any of our problems. 01:03:43.900 --> 01:03:47.700 And I think the thing is , like, the internet might limp along, it's pretty great. 01:03:47.700 --> 01:03:54.780 But I mean, to the extent that the internet facilitates human speech among 7 billion of us, that's a great thing. 01:03:55.040 --> 01:04:00.060 And when we take that away, I think we will miss it greatly. 01:04:00.180 --> 01:04:04.780 And so I think our final thought is, we need to be a lot more careful. 01:04:04.780 --> 01:04:08.460 We can't do legislation based on convenient bumper stickers. 01:04:08.460 --> 01:04:27.900 We really have to understand very carefully the dynamics of how law works, how human beings work, and how we start driving behavior when we start imposing punishments or even just uncertainty on people who are talking, expressing themselves, or facilitating that expression. 01:04:27.980 --> 01:04:29.780 Any other final thoughts? 01:04:31.720 --> 01:04:38.640 Yeah, I mean, just to say again, like, please take action, please call your senators, take action against both of these bills. 01:04:38.980 --> 01:04:41.640 There's a question about whether this would be bad for the US economy. 01:04:41.640 --> 01:04:53.300 I think the other bill, the LAED bill is a greater threat to the economy on the basis of a similar bill having been passed in Australia a couple of years ago that has really helped to screw up their tech economy sector there. 01:04:53.500 --> 01:05:04.560 So I think we would see, yes, a negative impact from the LAED Act, but also from EARNET, because of the lesser amount of vibrant services that we'd be able to see online. 01:05:04.800 --> 01:05:07.820 So anyway, I could rant about this all day. 01:05:07.820 --> 01:05:10.960 I can rant about 230, my favorite thing. 01:05:10.960 --> 01:05:13.320 I mean, I want people to understand it. 01:05:13.320 --> 01:05:19.360 I want, you know, right now, there's an awful lot of people with an awful lot of agendas saying, Section 230 bad. 01:05:19.440 --> 01:05:23.180 And I think the general populace is like, I don't understand. 01:05:23.180 --> 01:05:24.500 Yeah, that sounds good to me. 01:05:24.500 --> 01:05:29.100 I want everybody who's listening to be somebody who says, Aha, maybe that's not good. 01:05:29.100 --> 01:05:30.700 Maybe I should really rethink that. 01:05:30.700 --> 01:05:39.100 And to start, I think if people really understood what is being threatened by this legislation, they would not be for it. 01:05:39.200 --> 01:05:42.980 But I, you know, laws is kind of scary and complex system. 01:05:43.040 --> 01:05:45.760 And you leave it to other people to do that thinking for you. 01:05:45.760 --> 01:05:53.320 But I think, is make it that you can understand the mechanics of how this works, to draw an opinion. 01:05:53.320 --> 01:05:55.920 And I think that opinion will be that this is not a good idea. 01:05:55.920 --> 01:05:58.880 It's just that it's not always intuitive to understand this dynamic. 01:05:58.880 --> 01:06:00.020 How does the law work? 01:06:00.020 --> 01:06:02.120 Why do things respond to it? 01:06:02.120 --> 01:06:03.740 So hopefully, we can demystify it. 01:06:03.740 --> 01:06:17.340 And there's an awful lot of advocates out there, me, Rihanna, EFF, other organizations who, you know, are happy to sort of answer questions and help sort of make people understand what's really at stake with any of these bills. 01:06:18.900 --> 01:06:22.740 So I think that actually brings us to an end. 01:06:22.740 --> 01:06:25.480 Thank you so much, Village, for having us. 01:06:25.480 --> 01:06:27.620 And thank you, Rihanna, for doing this with me. 01:06:28.280 --> 01:06:29.500 I think this was great. 01:06:29.500 --> 01:06:31.480 And hopefully everybody watching thinks so. 01:06:32.720 --> 01:06:36.460 Great, thank you so very much to both of you. 01:06:36.460 --> 01:06:40.920 It was a pleasure to wrap up our very first day with you both. 01:06:41.200 --> 01:06:44.880 And so thank you, again, for speaking at the Village. 01:06:45.620 --> 01:06:49.140 Folks, this is the end of the day for us. 01:06:49.140 --> 01:06:59.000 We have survived day one of Virtual Crypto and Privacy Village, which frankly took more years off of my life than maybe all of 2020. 01:06:59.000 --> 01:07:00.380 No, I'm just kidding. 01:07:00.380 --> 01:07:15.840 But I will say, like an influencer, because I've never really streamed, if you like and subscribe below, we have various different avenues that you can communicate with us, reach out to us, engage with us. 01:07:15.840 --> 01:07:22.680 Since we can't be with all of you in person this year, please feel free to reach out to us on Twitter. 01:07:23.080 --> 01:07:28.380 For volunteers, I'm reaching out to you about a happy hour tonight. 01:07:28.480 --> 01:07:33.360 And to all of our speakers, thank you, thank you so much again. 01:07:33.360 --> 01:07:40.120 And if I haven't said this enough, thank you so much to everyone else who makes the Crypto and Privacy Village possible. 01:07:40.120 --> 01:07:52.760 There are a ton of folks who spend a lot of time, who have done so much work for the Crypto and Privacy Village this year to make it happen as we shifted gears from going live to going virtual. 01:07:52.980 --> 01:08:00.580 I could list all of their names now, but I'll save that for another time when I get to host and see you all again. 01:08:00.580 --> 01:08:07.320 So I'm going to transition and hopefully not break OBS because I'm doing both. 01:08:07.320 --> 01:08:17.160 So let's see if we can go to something that is a little more on five. 01:08:17.160 --> 01:08:19.880 Thank you, everyone, and I'll see you soon.