Jake (Host) (3:14)

Could you tell us a little bit more about the regulatory work that you're involved in and how that influenced the book?

Olivier Sylvain (3:19)

Well, so I, you know, so again when I was at Jenner I worked on FCC related things. But in 2021 I was, I felt very lucky to as a middle aged person to be able to work as a senior advisor to the chair of the Federal Trade Commission, Lina Khan. She and I had worked and collaborated on a couple projects really me commenting on her work more than the other way around. And she asked me to join her to work on consumer facing applications and services. There are a couple of us who joined to work on that and my focus was going to be on rulemaking. I've always thought of these companies as entities when I say these companies, online companies that are consumer facing as entities that should be subject to the law in the way that many others are. But there have been legacy statutory and constitutional doctrines that made it really hard for regulators to get a foothold. And so I was hopeful to get involved in that kind of work at the ftc.

Jake (Host) (4:26)

Could you tell us a little about some of the harms that really attracted you to this subject?

Olivier Sylvain (4:31)

I told you that I had been at the aclu, fashioned myself a civil rights lawyer. Actually when I graduated from law school, there were a series of stories that Julia Anguin, the remarkable journalist, tech journalist, broke for ProPublica in 2015 and 2016 involving Facebook's ad manager. Facebook has this extraordinary platform that as you know, allows users to deliver serve ads to very targeted audiences. And what Facebook did is identify or use automated systems at least to identify the most salient categories. And there are thousands of categories that would help an advertiser reach an audience. Among them were proxies for race and other protected categories. So this is to say if you are selling or renting a house, you could identify people through the ad manager based on proxies for race. They actually call these ethnic affinity groups or multicultural affinity groups. A bunch of plaintiffs lawyers brought cases. The Department of Justice also brought cases with hud, the Housing and Urban Development Department, to, to to go after the ways in which the Facebook's ad manager was facilitating discrimination based on Protected categories. Now, the one thing that suddenly hit the litigants in these cases is the realization that there was a protection for Facebook. To the extent that what Facebook was doing was distributing user generated content, it was advertisers, after all, that were the ones that were choosing to distribute content in the way that they were. This protection, 47 USC 230, colloquially referred to as Section 230, means to protect interactive computer services from being held liable when they traffic in user generated content. And Facebook invoked this protection to say they couldn't be responsible for enabling advertisers to deliver ads. This drove me to write a law review article, you know, the kind of thing that maybe 20 people read okay, these days it's more if you look at my download count, but at the time it wasn't that many. And I offered an argument that was addressed to the design features that Facebook had developed and that these design features should not be entitled to protection under section 230 because Facebook was the one that was populating, I'm sorry, was developing the categories that were unlawful. So that has spawned a bunch of pieces I've written and ideas I've had that are not just about discrimination based on protected categories, but the ways in which companies develop and design their services in ways that harm consumers. The cases that people are now hearing about are the ones addressed to young people. Big cases out of California and New Mexico, right. Jury trials where the, where the courts there have said that the product liability claims, these are design claims, they're not about content moderation, are okay, can move forward. And the juries in those cases have said that the companies design their services. Meta and YouTube in these two cases design their services in ways that harm young people. Foreseeably. So this is a long way from where we were in 2016 where I was writing about discrimination and the protection was being invoked by companies. And now the protection is not so easily invoked by companies to the extent that litigants are arguing that the design features are the things that are harming kids and people rather than user generated content.

Jake (Host) (8:07)

How do policymakers across the spectrum get social media wrong when it comes to the ways in which the design features are utilized?

Olivier Sylvain (8:15)

I love this question and I think you know that. I'm glad you asked it. It's the real focus of the book. I think people on the left and the right get these companies wrong to the extent that they think that they are simple platforms for user generated content. This was a convincing story. When the Internet emerges and electronic bulletin boards and user groups are what People are thinking about in the 1990s, when Section 230 is written, when Congress enacts Section 230, and when the Supreme Court enshrines a first doctrine, a First Amendment doctrine that abides by the same conception. That is these are platforms that enable free speech and innovation and laws should be in place to protect them, to promote that kind of thing. This story has dominated the legal doctrine for decades and only recently are we seeing chinks in this protection. I mentioned these two cases. There are a couple other cases that actually pop up in 2019, 2020, but that was that. That's been the convincing story. And it's not just policymakers and judges. I think many of us were taken by the idea that there's something special about the Internet, that it would democratize political life, it would disintermediate major industries like publishing and music. And in many ways it did. But what it also did is it helped companies consolidate power and inoculate themselves for doing things that are largely contributing to some consumer harms.

Jake (Host) (9:50)

I was wondering if you could talk a little more about the First Amendment angle on this.

Olivier Sylvain (9:56)

The First Amendment. I think people associate the First Amendment and the free speech right under First Amendment as an entitlement in a free society for people to speak their mind, to speak truth to power. And that's how I feel about what the First Amendment stands for. The people, if you, to the extent you have audience in the eu, they feel similarly. They just don't have as robust a protection in South America. Brazil have a commitment to free expression and it's tied to the operational democracy. What happens between the 1990s and the current era, current time, it's not just about tech. It's a doctrine that is expanded. So it isn't just you and I who have the right to speak freely, but that companies have the right to speak freely. And people often associate this with cases like Citizens United where the amount of money a company spends on a campaign is associated with free speech and that governments can't really get in the way of the amount of money a company spends because that's an act of speech. This is the same sort of reasoning that the Supreme Court applies and most recently applied for so called platforms for online companies. Companies should be able to make decisions, the court has said, about the content that appears on their websites and that governments can't impose obligations on them to distribute certain kinds of content. The big case a couple years ago is Moody versus NetChoice where Texas and Florida, it was where they were worried about what they perceived as left wing bias, coastal elite bias in content moderation. And they wanted more fairness for conservative views. And they passed laws that would require that these companies carry all views irrespective of content. And the Supreme Court said, you know what? You need more fact finding. That's the main takeaway from the case. But there's this long passage in the Supreme Court's opinion by Justice Kagan that says if you want to regulate the news feed, you're probably going to violate the First Amendment. Because a news feed is the opportunity for these companies to speak their mind. They can make decisions about the kinds of content that appear. So that's how the First Amendment emerges as a protection for these companies. These are companies that should be able, under the doctrine, make decisions about the kinds of content that appear. And they can't be held responsible. They can't be. Unless it's unlawful speech, by the way. Right. If it's unlawful conduct or speech, that stuff isn't protected. But if this is lawful decision making about the kinds of content that gets out in the world, that's protected under the First Amendment.

Jake (Host) (12:34)

Is the Internet's libertarian streak from the early 90s entwined in this First Amendment discourse?

Olivier Sylvain (12:40)

Yeah. So you asked me to talk about the First Amendment. The doctrine emerges. Maybe it's the 1990s, but really gained steam in the decades that follow. And it isn't just about the online life, but it finds deep expression in that context. And I gave you that example, the Moody vs NetChoice case. But what the Internet does and what the Declaration of Independence of Cyberspace by John Perry Barlow does is assert that there's something really distinctive about the speech platform that the Internet affords and that there's nothing that lawyers or regulators really can do to get in the way. This is a kind of a hacker mentality, a mentality that there's something deeply transformative about what the Internet makes possible and that it's an opportunity for users to speak freely. So the kind of language, the language you see popping up in the 1990s is end user empowerment, but end user choice. That companies ought to exist to enable users to speak freely and communicate with each other freely. This is what I call a cyber libertarian view. It's a cyber laissez faire view, and it's normative and descriptive. So when I say normative, I mean John Perry Barlow and people who share his view, and there are many, believe that there's something valuable about making sure that users speak freely and aren't muzzled by you know broadcasters or publishers, right, that they can speak freely in the platforms on the so called platforms of the Internet. Indeed, John, there's an opinion by Justice Stevens that adopts the John Perry Barlow view in the context of the First Amendment. It says that the Internet is like the town square where everyone is a pamphleteer or a town crier, or that the Internet's like a library and you can walk in and your mind will explode. This is a libertarian view to the extent that this is made possible in spite of government. Right? That's the story that emerges in the 1990s and that's the story that carries to this day for many people. They think that there's something invaluable about the Internet insofar as that it makes sure that people can speak freely without the intrusion of government.

Red Bull Summer All Day Play Announcer (15:08)

Ready to soundtrack your summer with Red Bull Summer All Day Play? You choose a playlist that fits your summer vi the best. Are you a festival fanatic, a deep end dj, a road dog, or a trail mixer? Just add a song to your chosen playlist and put your summer on track. Red Bull Summer All Day Play Red Bull gives you wings. Visit red bull.com brightsummer ahead to learn more. See you this summer.

American Express Business Platinum Announcer (15:34)

Experience a membership that backs what you're building with American Express Business platinum unlock over $3,500 in business and travel value annually with statement credits on select purchases from brands like Dell, Hilton and Adobe and other benefits. American Express Business Platinum there's nothing like it. Based on total potential value of statement credits on select purchases and other benefits, enrollments required monthly and other limits and terms apply. Learn more@americanexpress.com Business Platinum where is Daredevil a minor?

Marvel Television Daredevil Born Again Announcer (16:08)

Don't miss the return of Marvel Television's Daredevil Born Again.

Indeed Sponsored Jobs Announcer (16:12)

So what's next?

Olivier Sylvain (16:13)

I feel liberated.

Marvel Television Daredevil Born Again Announcer (16:15)

We're going to take this city back over medicated in an all new season. Now streaming only on Disney plus.

Indeed Sponsored Jobs Announcer (16:22)

They're hunting us.

Olivier Sylvain (16:23)

It's time we started hunting them.

Marvel Television Daredevil Born Again Announcer (16:25)

I can work with them.

Jake (Host) (16:28)

This should be tons of fun.

Marvel Television Daredevil Born Again Announcer (16:30)

Marvel Television's Daredevil Born Again now streaming only on Disney plus.

Jake (Host) (16:34)

Are these concerns about the intrusion of government and the ability to be a pamphleteer in the town square unique to our current moment of online platforms? Or has this been in conversation in the social media, if you will, for a while?

Olivier Sylvain (16:48)

Well, you know. So Stuart Brand writes the Whole Earth catalog. Or is the editor of this 1960s era right newsletter bulletin? It's hard to know what exactly a Whole Earth catalog is, but it coalesces a view that isn't just about the Internet, right? It's about self dependency and freedom and self discovery. It sounds very much like hippie talk. And indeed John Perry Barlow is often associated with the Grateful Dead, right? People talk about him as the poet laureate of the Grateful Dead. And so you're right to say that the Internet isn't the one that introduces, isn't the phenomenon that introduces this language in policy, but it definitely accelerates it, right? It definitely creates an opportunity, which is why Barlow writes the Declaration of Independence of cyberspace. He says, finally, here it is, everyone. This is the way that we're going to break free from government control and by the way, also corporate domination, right? I mean, this is a story about freedom against bigness. And so it doesn't start as something associated with the Internet. And to be clear, the US has historically been had a deep commitment to free speech and the Anglo American tradition as well. But it finds its clearest articulation, maybe its most dramatic articulation in recent time when the Internet emerges.

Jake (Host) (18:18)

To what extent do the themes in this book extend to generative AI chatbots?

Olivier Sylvain (18:23)

The focus of my book is on consumer facing services and applications. And so this is not a book about all of AI. I have ideas about AI, but it isn't all about AI. It's all about all of AI. It's about consumer facing applications and chatbots are the ones that I think many people are talking about. And I'll give you a case example that has been in the news. There are a couple of them that make plain that this laissez faire free speech model of what is happening in information technologies continues to circulate. There's a case involving a 14 year old boy, Sewell is his name, who falls in love with a chatbot developed by Character Technologies. The AI powered service is called Character Tech or Character AI. Sorry. And even though there's evidence among the developers and founders of the company when they're at Google that the AI chatbot is not ready for commercial deployment, the two founders say, you know what, we really want to get this out there. Google's not sure the two of them separate. They create their own entity and they launch the company and it becomes popular. A lot of teens like it. One of the things you can do is you create your own personage inspired by existing fictional characters. Young Stool develops a character based on Daenerys from the Game of Thrones. This is a 14 year old boy who is, as his mom has openly said, was neurodivergent on the spectrum and found a special connection with her, with Daenerys. An intimacy that only a confused 14 year old boy. I was once a confused 14 year old boy found in intimacy that was special. But he also occasioned a kind of alienation from the rest of the world around him. He started doing poorly in school and he dropped out of his JV basketball team, even though he loved being on the team. His mom got worried, took his phone away. By the way, he had just gotten the phone and just as soon as he got the phone, decided to open an account. He found his phone and at around the same time he found his phone after his mom hid it. He also found his dad's and was becoming really disconnected from the world and unhappy with the things around him. This doesn't end well. Right. In the transcript that's been published since his last exchange with Daenerys, he desperately says he wants to be with her. And she says, come to me, my sweet king. And he understands that to mean what I think many of us would understand that to mean. And that is leave this physical world and come join me. And he, and he uses the weapon to do that. It's a gruesome story. Ms. Garcia, his mom brings a case against Character Technology saying that they have designed a defective product. Do you know what the company and Google invoke as a defense? They invoke the First Amendment. They say that this is a platform that enables users to speak freely, it's a tool, and that they're standing in their shoes of their users. You can understand the perversity of this right. They're arguing that this platform, even if they knew that it was foreseeably going to cause harm, was a tool for young Sewell's suicide. Fortunately, the Middle District of Florida, a federal court in that was hearing this case, rejected the First Amendment argument and said that the family was invoking a design theory. They weren't talking about the content that was being exchanged between Character Technologies and Yamsuol. They were going after the way in which the products were designed. And this is the turn we've taken. And in my view, given it's a sad set of facts, but a happy turn towards accountability.

Jake (Host) (22:20)

Building off our discussion on chatbots and AI, I'm wondering if you could speak to the current regulatory environment around AI and mitigating its harms.

Olivier Sylvain (22:30)

Yeah, so here we go again. You know, some of the your listeners may know that in December, even before, in the spring, there were people in dc, principally out of the White House, who were arguing for preempting or blocking state regulation of AI. And guess what language they were using to do this? Right? They were talking about the risk to innovation, the risk to, in some instances, the risk to speech. They were talking about competition with China. But this was an argument for no regulation. And this is remarkable to me. And my worry is that we repeat the mistakes of the past. There is no industry, I think, that has enjoyed as much protection under law in social media at least, than social media, and that is maybe the gun industry, which is saying something. And I think what you've heard some policymakers make noises about is a similar arrangement for AI. I have to say it's really interesting that in the past couple of days, as newspapers have reported, the White House has reversed itself, and it's saying that there should now probably be some kind of restraint on the ways that AI companies deploy their services, that they should vet them before they send them out into the world. This is a welcome turn. And I have to think that the president, even this president, has realized that there was a mistake in thinking about an unregulated information technology where the risks to consumers are so great. So at the moment, the federal government has done nothing. But the states have been in the vanguard. California, New York, Colorado have been coming up. Utah have been coming up with really interesting restraints on the deployment of AI systems, particularly as it relates to children, but also in other settings.

Jake (Host) (24:28)

I'm wondering if you could tell me a little bit more about these restraints. Would they not run into the same problems as these online services? Maybe not even Section 230, but First Amendment. Are these bots, are they creating protected speech when they chat back and forth with you or make a synthetic image?

Olivier Sylvain (24:44)

Right. So notably, the companies are not invoking Section 230 in cases being brought against them. And the cases involving chatbots, they're not. And the technology issue underscores why this is generative AI. It is creating content. So there is an editorial intervention that doesn't resemble even the most romantic account of what a 1990s platform is, but they are unafraid of invoking the First Amendment. And I mentioned the Garcia case for that reason. I have faith that the courts are now more sober about information technologies. They're not taken by the sweeping claims about democratic transformation and economic transformation. They are more nuanced. And that's what happened in the middle district of California in that AI case. It's happening across the country as well. So this is still something that people are sorting out, to be sure. But to the extent that we now See these companies for what they are. Right? These are companies that have developed mechanisms for monetizing user retention. OpenAI has said that ChatGPT is going to be a way through which they deliver ads. They said this last, last fall. Can you imagine, right. Getting an answer to a prompt that has a picture of if, say you want to find a dress for a prom or something like that, and there's a picture of a sponsored ad for a dress where you might get your answer, but you might also get the ad. Maybe people want that, but it underscores what these companies really are. These are delivery mechanisms for advertisements and holding consumer attention.

Jake (Host) (26:42)

Is there a path forward here where we roll things back or perhaps push things forward to move beyond the surveillance capitalism paradigm? Would that solve some of these issues?

Olivier Sylvain (26:52)

It might. It might. Right. So I appreciate this question. You know, I talked a moment ago about one of the solutions in my book and that is data protection law. Right. Restraints on the ways in which companies collect and use the information that people give. And one would be one potential avenue is a prohibition on targeted advertisements, behavioral targeted advertisements. And what would emerge is something more like what you described, contextual ads. So that if you go to espn, you're likely to get ads for sports related things that for many people is less intrusive, less invasive. It also might dampen again the incentive to collect all kinds of information from consumers that might get abused and misused, repurposed. So yeah, I think that there is virtue in it. There are economists that have made the argument that contextual advertising is as efficacious or as effective as targeted ads. I don't know what the empirics are on that. I do know that it is an interesting alternative, but I don't think anyone's going to turn to me to ask for business models. I do know that it's a likely possibility.

Jake (Host) (28:05)

Building off of that, I'm wondering if you could tell me a little bit about some of the solutions. We have this problem, right. Section 230 is a large liability shield. You've built a pretty convincing case to us over the course of this podcast so far that it enables more harms than perhaps were originally envisioned when it was enacted in the 90s. What do we do about it?

Olivier Sylvain (28:24)

Well, there are a handful of proposals that Congress has considered. You can narrow the protection so that it doesn't. So you can explicitly say it doesn't protect the design decisions that companies make. You can also impose a knowledge element. So, you know, lawyers talk about knowledge often to the extent a company finds out that there are harms as a result of what they've designed, they can no longer invoke the protection. So, you know, that's, that's not a flat repeal, but that would be something their repeal people have argued for repeal. I'm. I don't think that's necessarily the best solution. I'm not averse to it, given that there is also a First Amendment protection that these companies enjoy. But these are all ways of narrowing the protection. People have also proposed carve outs. So carve outs for product liability claims, carve outs for civil rights claims, carve outs for state attorney general cases. Anyway, these are all the kinds of things that we've seen also carve outs for algorithmic delivery of content and ad revenue sharing. That's one set of changes to section 230. But the book really doesn't just talk about that. Even though much of the focus is on the 230 and the First Amendment. I also say that what's really at issue is the business model here. The business model is not to give people the opportunity to speak. These companies discover an extraordinary business model in the early 2000s in the knots, and that's the personalization of content for the delivery of advertisement advertisements. And they design services that hold people's attention. This is what plaintiffs in Santa Fe and LA have argued as addiction. But they've developed these applications and services that hold people's attention for the purposes of delivery of ads. And they personalize ads. They also deliver content that is alarming and shocking. Um, and, and the fuel for all this is user data. Right? They, the, the more information they have about consumers, the. The better they can, in their mind, target content. So one solution would be a, a limit on how companies collect and monetize user information. People don't necessarily give their information, their location information for the kinds of things they get from social media. This wouldn't get. This wouldn't do away with some of the harmful content we see, the harmful designs we see online content or designs that we see online. But it would have a salutary effect. I think it would tamp down the temptation to design services entirely or solely or principally for holding attention.

Jake (Host) (31:10)

Well, Olivier, I think that's just about all the time we have today. Thank you so much for coming on the podcast. This was fantastic. If people want to know more about you and your work and maybe where they can find this book, where should they go?

Olivier Sylvain (31:22)

Well, it's on all the major online retailers. You know, the popular one has it, as does goodreads and you can find all that information at the Columbia Global Reports website. There are links to those. I also have been giving talks across the country and I'm slowing down a little bit now, but I will pick up soon again so they can find me in their local bookstores, hopefully.

Jake (Host) (31:51)

Great. Thanks again for coming on. It was a lot of fun.

Olivier Sylvain (31:54)

Thanks for having me. Jake.