Welcome to Tech Talk. Bye. Tea. Tea.

Thank you so much. And Nicole, I was so thrilled when you asked me to have this conversation.

Better than me being up here alone. So No, no, not at all. There are a few,

magazine articles that deserve ten year anniversary celebrations, but this was my favorite article ever. The one I thought was most interesting because ten years ago, through with your guidance, we put our fingers on the central challenge of free speech at the dawn of the Internet age. Namely, what to do when lawyers for private companies, and you were then, Deputy General Counsel at Google, are more powerful than any king or president or Supreme Court justice when it comes to who can speak and who can be heard online. So what I wanna do in this really meaningful conversation, meaningful because I know we're gonna shed some real light, is begin by identifying what was the problem we were talking about ten years ago. Then talk about what is the problem today and how is it different. And then begin to think about some solutions. So when we talked, you told me about your struggles with Thailand. Tell us about what Thailand was doing in Google and what your response was and that how that exemplified for you

the challenge of the content decisions that you, as Google's decider, were being asked to make. Yeah. So first, thank you for traveling down Philadelphia to be here. You're welcome. I say, like so in this article, Jeff raised this issue of, like, Nicole's the decider as called in the legal department at Google because she decides all these content issues. A, it was never just me. B, it was a joke propagated by Alex McElder, who's sitting in the back, who was working with me at the time. And every time my children hear this whole idea of, like, I'm the decider, they're like, mom, you can't even make a decision on what vegetable we're having for dinner. Are you kidding? But at the time at Google, one of the things in the mid two thousands, that we did was we we had acquired YouTube, which really changed the landscape for how we had to deal with multiple countries in part because it was visual. So instead of just having text, in languages that some people couldn't read and some people could, you suddenly had this medium that was instantly global because the pictures would tell it all, but they were understood in different countries in different ways. And Thailand was one of them. So in Thailand in 2006, a number of videos that were critical of Thailand's king, went up on the internet. There were depictions of him with like a foot on top of his head, which is very offensive in Thailand. Pictures of him as a monkey, and other things. And, in Thailand, it is illegal. They have a Les Majestay law which makes it illegal to criticize the king. And, And there were a whole bunch of things that had to fall into place for us to deal with that. One is to understand the law. The second was to understand sort of what videos were at issue. The government had contacted us about 20 different videos that they felt were violating their law. And then there was the hardest decision about what is it we're gonna do about it. At that time, we kind of decided, well, maybe we won't do anything about it. We'll just leave it up. We don't have anybody based in Thailand, so do we really care? And I eventually went to Thailand, and sat down with both the government but also with people who, were part of internet associations and internet companies and civil society in Thailand to understand the problem. And one of the most important conversations I had was actually with our US embassy officials in Thailand who explained to me, they're like, this is a country that has had 21 coups in the last thirty years. The king, who is now 80 years old, is the only stability they have ever known. And so his stature in that country is something akin to George Washington, Elvis, and Jesus Christ all rolled into one. And all of a sudden, like, a whole bunch of things made much more sense to me about why we were getting this reaction and it was not just from the government, but it was from the people themselves, who were really concerned about this content. And that, for me as a First Amendment lawyer, changed a bunch of the ways I thought about what our responsibilities were as a global platform. In the end, we decided to respect the democratically created law around Les Majestay by blocking only for Thai IP addresses access to videos that were clearly in violation of the Les Magistae law. That's kind of a narrowing of what the original ask was. I will tell you, and I I feel like sometimes you make these decisions and everyone's like, well, that's the answer. I am still deeply conflicted about whether that was the right answer. I am deeply conflicted about whether IP blocking is the right solution for us on global content. I am deeply conflicted about what the parameters are that we set for deciding obeying one country's laws versus another. And I don't think now looking back ten years, right, there's a bunch on the Internet that has changed, and yet the difficulty of some of those decisions, I think, is is the consistent thing, which is that they are always hard, and and will probably always be difficult.

Wonderfully summed up. So ten years ago, one of the central problems was what the content policy should be and how to apply them locally and globally. After that decision, there were calls of global IP blocking by some countries, which you nobly resisted and, in some cases, were shut down in places like Turkey as a result. So there's still a debate about content policies, and Facebook, as well as Google and Twitter, are facing increasing pressure to ban ever more speech that the First Amendment protects. The European, court has imposed its right to be forgotten, which is much less First Amendment friendly than things were back then. So, describe how the question of content blocking looks different ten years now than it did then, and how the pressures seem to make it harder for the companies to basically embrace the First Amendment values that you were trying to apply. Yeah.

It's interesting, like so ten years ago I was just, like, looking back at some of the stuff that I was doing ten years ago. Ten years ago, I was testifying before the senate on why we were withdrawing from China because of the increasing censorship in China, because of the hacking of Google's systems and its targeting of Chinese dissidents on its platform. And the and the conversations were around, like, how can you be in in markets like that, that are so authoritarian and and censoring? There at the time, Google was blocked, had been blocked in 25 different countries. And we thought that that was terrible, right? And it was YouTube and Blogger and Orchid, which no one remembers, but Google had a social network called Orchid and it was in, like, three countries and blocked, in many. But the, like, at the time, we thought that the biggest threat to democracy and the company's mission and our ability to deliver service and access to information, we thought the biggest threat was to be blocked in a country. That content could not be made available. That information could not be be be discovered in those countries. And now, ten years later, that feels really naive. I think the thing that I did not foresee is that our openness would be exploited, and that these platforms we created would be weaponized. And I I think that that's so now when I look look at some of the congressional testimony that's happening now with Mark Zuckerberg or with the the array of social media platforms that had to testify about terrorism recently, and the questions are questions like, why are you allowing that content on your platform? Why aren't you getting it down faster? Why aren't you verifying the identity of the people who are there and turning them over to law enforcement? Those were all questions, by the way, that if it were done by China, we were gonna resist strongly. But now those are the questions that our own government is asking. Why aren't you doing this more and faster? And and it's happening in places like Germany, right, and The UK and and and others. And,

I think that it is not wrong to ask what the what the solutions are to the weaponization of the platforms, but I'm very concerned about the direction of the where where policymakers are taking it. That's fascinating. Ten years ago, we were talking about the British trying to remove bullying content and senator Lieberman trying to remove some terrorist recruitment videos, but you're now saying that democratic governments like Germany and The US and Britain are demanding the removal of far more speech than they were ten years ago. Is there any hope of encouraging the companies to follow the first amendment in the face of this consumer and government pressure? And if so, how would you shore up their ability to follow first amendment pressure? Well, the other countries are not gonna follow the first amendment. One. But

No. No. No. Of course not. You're right. But but article 19, right, of the Universal Declaration of Human Rights, the the the human fundamental human right to free free expression, there there is that, and and we should hold them to that. And if you haven't read, David Kaye, the the UN special rapporteur's report on on freedoms in social media, that's really well worth reading. So so I think that there is a backbone framework for all of the, I'm gonna get this wrong, a 172 countries who signed on for the Universal Declaration of Human Rights and and some who who signed other similar types of agreements. There's a backbone to help make these decisions. I think that that we're we're in this really fraught time of thinking that the solutions are about removing more content. And that to me is a losing game at the end of the day. We can keep chipping away at this, but the the honest answer is there has always been bad destructive content. There will always be. And the platforms, whether it's a newspaper or a social media platform, is going to make mistakes. And and so now that I sit outside a company, I get to say things like this. We make mistakes. Right? We do not get it right. We either take down too much or too little. And what you try to do is either make sure that you're not making really bad mistakes or that you have a rescue plan if you did make a mistake. And and so I'm not I understand that there has to be a conversation around some sort of restriction regulation. I'm just not sure that content is necessarily the right place. Why aren't we looking harder at competition law? Right? Why aren't we looking at at making sure that we're not fully dependent on these singular large platforms and instead encouraging the creation of what I understand Ethan might have talked about or or I'd heard in a couple other, rooms, right, other experiments in social media that are alternatives to to the current places that were were captured.

Okay. So one big problem which has changed is the question of content standards and content removal, and you've said that there's more pressure than there was ten years ago to remove content that might be protected by the First Amendment. There are at least two new problems that, folks have been talking about this morning, and they include, disinformation or fake news and polarization. And they relate to the fact that our discourse is now operating at a warp speed that the American founders would have feared, and we're being governed by Twitter mobs and passionately shared Facebook posts that can either be based on disinformation or can increase polarization. Describe how would you put these two new problems, and to what degree are they new? Yeah. So so I think,

let me step back a little bit on sort of just the design of the platforms. And and taking, like, Google search as as, like, one marker. Right? When when in the design of Google search, the the engineers were designing around certain pillars of of of value, which is we wanted comprehensiveness, relevance, and speed. Comprehensive enough to get as much content as we could, right, to fill the library. Relevance to be able to deliver the right answer at the right time based on your question. And then speed to get it to you as quickly as possible. That all made sense for search. At a point in sort of the mid two thousands when you had social media and this notion of the social graph on the rise and personalization for the targeting of advertising, so behavioral advertising went on the rise around the mid two thousands. Those combine to change the pillars of design in the new platforms, which we now call social media. Right? Which is engagement is the metric of success. How how much time do I spend looking at various things on your platform instead of going off and finding it elsewhere? Personalization, which is not the same as relevance. It's not does this answer your question? It's here's more stuff you like, and and speed stays. So so on those pillars, you have this engine for a rocket ship based on the most outrageous content that we can serve to you that keep you here for the engagement, and increase as we've I've been hearing some of the the research being done right now, but increase the polarization because of that, because of the the way we act as humans when when we communicate with each other. I think that those pillars are super dangerous right now. And and and the and we've kept this notion of speed that everything has to go so fast. You have to have more and more and more all the time that you're engaging with. What if we say no? Right? What if we as a society say, like, that's not the way I wanna consume information anymore? For me, and I've said this publicly in a couple of forms now, like, for me, I would prize a platform that was designed around authenticity of the content that I see. Not necessarily that I know exactly the person who who, generated it, but like where where did it travel from? What's the provenance and and and the the trustworthiness of a piece of information? What's the accuracy of it by some metric? And and what's the context for understanding what I'm seeing? What if we took the incredibly talented engineers who are working on ads optimization and moved them to that? What would we end up with instead? And and I think that that's we are now at a point where that's the self reflection the companies should be doing.

This is a huge insight. You're saying that the platforms themselves designed to promote personalization, relevance, and speed may be creating it's really a Madisonian problem because Madison believed that the whole point of the constitution was to slow down public discourse so that reason rather than passion could prevail. And, through the design of the platforms, we're finding a Madisonian nightmare where posts based on passion travel faster and further than those based on reason. So you're if you were designing a slow speed movement for the Internet to slow down deliberation and promote Madisonian values, what are some of the philosophical and technological features that it would have? So first of all, I love talking to Jeff because he makes me sound so much more smart than I am. No. That's not true. That's not true at all. These are you are, in addition to the decider, the chief is Madison of the internet. So someone's gotta do it, and it's gotta be you. I gotta take that one back to my kids. It's yours.

So I think I think some of the things that would be interesting, providing more context for the content that we see would be really interesting. They would also be is when I think about trying to bring that to, like, an engineer, they'd be like, that just causes more friction for the user. That's getting them off, like, the goal. And I'm like, yeah. But the goal may not be right anymore. So, like, why don't we try that? What if so so there's, a group, led by Debroy and Eugene Yi at MIT doing a at Cortico doing some projects on visualizing how discussion happens. And I was just looking at they have this, like it's like a tree. It's like one tweet is like a green leaf that comes up on, hey, I'm feeling x. And then they grow the tree as the conversation grows with red leaves being sort of, contrary or harsh or hostile content, and green being, like, more sort of supportive con conversation. What if every conversation we had, we could track, how is this conversation going? Is it getting more hostile or less hostile? Because one of the things that we're missing in in the platforms we have is the social cues that tell us when we sit around a table, we look and see someone's face says, why would you say something like that to him? Right? We need something equivalent to that on the platform. So what if we we sort of prioritize that as like, how do we create more social cues for each other about how our conversations are going? That would slow us slow us down, but it might be really interesting. I also think I've been toying with this idea, so forgive me because it's such a nascent idea, but what if, instead of what some countries have opted for, which is a a media blackout before an election. So you saw that happen in France recently. It happens in Brazil, certain number of days before an election. They basically just close off the media. Candidates cannot speak to the public. I find that really distressing and, frankly, not very effective because then, like, all the shenanigans just move before the window, so that doesn't really help. What if instead we stood up a neutral body, the FEC, Pew, I don't know who it would be, but some group, that all of the media platforms would dump their content in force x not forever, like, not to, like like, reveal all of your IP, but, like, literally just for some period before an election so that we could track where information comes from. Did it originate in a four chan conversation? Did it did it travel through through Breitbart before it made it to Twitter? And and and then ultimately into the mainstream media. I don't wanna censor any of it, but what if we could just know where it came from so that we would have context to understand how we evaluate it? That would be super interesting. That, I don't think, is where you necessarily would wanna put your engineers right now because there's no incentive to do that. So I think there's a there's a large public policy question about how do we change the incentives of what's getting built, and and where does what kind of regulation does that look like?

What about the role of the editor in chief? You were the decider for Google and found it challenging to make all of these content decisions. Is it, and and Facebook has contemplated prioritizing on its news feed stuff that people actually read. So some data suggests that fake news travels more quickly than real news because people pass along ex inflammatory headlines without reading them. Today, I think there was some evidence on the other side. Do we need the Google and Facebook and Twitter deciders to be deciding what people should see in order to provide context rather than just what they want to see? And what would that editorial function look like?

Yeah. So so this is where when you're not in a company, you get to say, I don't know the answer to that, but that's a really interesting question. So so here's what I think is is very hard about that. I don't think I trust a private company any more than I trust a government to decide what content we see. So, I'll start from that place.

I'll I'll talk to that one as well.

But I do think we need which which tells me this, like, then then this is not just a solution from the tech companies. Right? We need to engage all of society in how to fix this. One of the problems to me if you if you look nationally and then it's exacerbated when you look globally, right, is even if you're a tech company and you want to get a diverse array of content, it's not always there or you don't necessarily know how to find it. So so I think in like, for example, in this country, I think the demise of local news has had a catastrophic effect on our politics. And so all of the conversation exists at this national level, which is highly polarizing as opposed to, hey, locally, I care a lot about whether that bridge is getting built or my water is clean or my kids are going to the right schools. And and those are places where community can be built, but we don't have anyone on the ground writing that news. And so the question for us as a society is how do we encourage the development of more of that local news so that we can feed that into the tech Twitter streams or Facebook streams or or or YouTube streams to counter some of the others. And and what I worry about is that we've just lost some of that conversation, and that's on us. Right? That's not the tech companies weren't responsible for also building local news. It is on us to ensure the success of that type of content. That is exacerbated, frankly, at a at a global level where it is really hard for a company sitting in Mountain View or Menlo Park or Seattle or wherever to decide what's the trustworthy news and the widest array of news on the ground in a country that I have never been to or may not be permitted to be in? How do I surface the right information from that? How do I find trustworthy sources? And and that is really, really difficult. And that is a global project for us then, right, which is to ensure press freedom globally. It is to ensure that what we had, I think, ten years ago believed we could be, which is to be the voice for those who are unheard, somehow, as societies, we actually need to be having more participation at that level.

The idea of shoring up local news is a wonderful Brandeisian solution. Louis Brandeis is a hero of both of ours. He denounced the curse of bigness. He thought that only on local levels could reason thrive. And by prioritizing local news, the platforms would create a market for them, which could be a virtuous cycle. Facebook has proposed creating a Supreme Court for Facebook. What do you think about the idea of a Supreme Court for Facebook? Should they create it? And if so, what should it look like from the appointment and nomination and confirmation procedures to the substantive law they apply to the, to to to the possibility of, an appeal.

Yeah. Okay. So, like, I'm not gonna comment specifically on the Facebook thing. Kate Klonick and, I think it's one other author who I'm currently forgetting just read an op ed, which I thought was really good. I'm not gonna dismiss any possible solution. Right? Because we clearly there is no single silver bullet for this, so we need many different solutions. And I think the notion of an independent, I'm assuming it's independent, third bot body that makes this sort of oversees or or provides advice on whether you've made good or bad content decisions, that seems to answer the mail on one of the key criticisms for these platforms, which is about transparency of their decision making and accountability for that. And so it matters very much then, is it independent? Who's paying for it? What's like, who's getting appointed and on what basis? And does do we really expect some panel of, like, could it be larger than 50 to make decisions for the world? Because that's a lot. Right? But Facebook I was just talking with Elliot Shrey. They've got 30,000 people working on content, issues. So, like, how does a small panel of a court really manage all of that? I I think there's lots of detailed questions, to to deal with in in in that. But but we should try. Right? Again, and I I think that that's an accountability and transparency answer, but I don't think it gets to you can't have, like, every piece of content going up. How many like, it's hundreds of hours going up on YouTube per minute at this point. Like, not all of those are gonna get seen by a court. Right? And so you're still gonna have the issue of content that comes down, and at the decision of private companies or their contractors. You still have to wrestle with those choices. And I I also think I I think you still have the global problem. Right? Which it's really hard to capture global nuance when you're a singular body, and and I'm not sure it's even appropriate to have some singular body wherever it is they said Geneva or whatever, making decisions for very local, communities.

Last question and last words to the audience. What's so exciting about our conversations, all of them, is that you are challenging us to think about the Internet in constitutional terms. Not legal ones, but philosophically, you're challenging us to resurrect the frictions, or the cooling mechanisms, as Madison called them, that will allow in the online space the same, reason deliberation that the Framers thought was necessary on the internet space. You and I have talked about, the possibility of the National Constitution Center, along with many of the organizations in this room, coming up with a project that would apply this constitutional thinking to create a slow speed movement for the Internet and would try to identify philosophical and technological solutions for promoting Reason Online. If you were going to design a charter or mission for this project, what would it include, and why is it important for everyone in this room to join us?

Oh, that's a big question, Jeff. I think, I don't know that I have a charter in mind. I do think there is a uniqueness to these social platforms, particularly given how dominant they are in in our discourse. Right? Where, we need some greater guidepost than we've had to date. And and in service of democratic values, and human rights frameworks. And and so, how we institutionalize those and propagate them and and get people to design for them. Like, we can do all the research we want. I think I would what I heard, like, some incredible social science research is being done now on, like, what is useful conversation? What is a public a healthy public sphere? All of that is so useful, but then the question is, how does that become a design? Right? And how does it become either a business model or to Ethan's point, a public service, driven model that that we can support? And but I think the first point is, like, agreement on the principles of what we're here for.

Ladies and gentlemen, for her contribution to promoting reason on the Internet, please join me in thanking the decider, Cohen.