与 Sarah B. Rogers 讨论欧洲言论自由失灵与“审查工业复合体”

David and I are staying in a 300-year-old house, and we’ve both smashed our heads on the beams twice already. This is both our first Davos, David. We’ve been here for 24 hours. Any first impressions?

Yes, first Davos. It’s interesting. We’re staying very far away.

They apparently didn’t want you to be part of this. They didn’t want me too close.

They didn’t want me too close.

But we finally got you an invitation. Your invitation did not get lost in the mail.

My invitation didn’t get lost in the mail this time. But those of you who watch the pod know what I’m talking about.

Inside joke. Yes. But it’s great to be here at USA House. Thanks to all the sponsors. We’re really delighted to have our first guest on the pod, Sarah B. Rogers, who works as the Under Secretary for Public Diplomacy at the State Department.

For members here who don’t know this position, or what you’ve been charged with, or what you’ve decided to work on, I’m curious about that. Did they tell you what to do, or did you come up with your own mandate? Tell us everything about what you’re doing.

Longtime listener, first-time guest. Thank you to both of you, and thank you from all of us at America House for joining us here.

I am the Under Secretary for Public Diplomacy. When I got this nomination, my friends and family all congratulated me and then furtively said, “What is that?”

Diplomacy traditionally concerns the relationship between the American government and foreign governments. Two ambassadors shake hands, make a deal, solve the war. Public diplomacy is different. Public diplomacy addresses the relationship between the American government and foreign publics.

This has become a very important undersecretariat with the rise of the internet and, especially during the Biden administration, these mushrooming concerns about so-called disinformation and what we do when there are allegedly malign influences on the public’s view of America, and on the intersection between the public’s interests and American interests. How do we interact with the internet and the information ecosystem? That is part of my portfolio.

I also oversee other soft-power activities, including our educational, cultural, and sports diplomacy. I am privileged to play a role in the World Cup this summer and the LA Olympics coming up, as well as the Fulbright program and others like it.

You seem particularly focused on freedom of speech and a little bit of tension between our standards and the companies in America that have made the move to strongly support freedom of speech—something that got lost in our industry for a couple of years in technology but has now made some significant progress on.

It seems like some folks in Europe don’t share our love of freedom of speech. Maybe you could explain to us what the tension is today and what some of the regulations are that have been put in place in Europe.

Sure, absolutely. The 2 main regulations that I’ve interfaced with since taking office—part of this is just a product of my first official trip being to Europe. While I was in Europe, a large fine came down on an American platform, X, under the Digital Services Act, which I’ll get into in a moment.

Europe, especially since the Second World War, but really since the American founding and our codification of the First Amendment, has taken a much stronger approach on free speech than even most of the West. With the rise of the internet and all, or a lot, of communication becoming transnational, we see these new technocratic regulatory frameworks in Europe bumping up against the commitments to free speech in the United States.

Jason makes an important observation: For a while, some of these large American technology platforms were more inclined to moderate or censor in conformity with prevailing norms and concerns in the United States. But I think in the United States we’ve shifted back toward a less censorious approach, and so have these platforms.

At the same time, you have regulatory efforts in Europe and the UK. I’ll name a couple that I think have been particularly relevant. The UK has something called the Online Safety Act. The Online Safety Act imposes age-gating obligations on a broad swath of content, almost any content that’s upsetting. It then requires platforms to run risk assessments and, in some cases, remove content that the UK would say is illegal.

In the UK, major categories of content are banned and rendered illegal that would not be illegal in the United States, which is where these platforms are located, where their original user base is, where their executives live, and which is their primary regulator.

Under the Online Safety Act, we now have active litigation by the relevant regulator, Ofcom, against several American websites. These are websites that don’t reach into the UK. They’re not websites dedicated to discussing the Queen. They’re not websites that sell goods in the United Kingdom. These are websites that exist on American soil, host large quantities of American users, and oftentimes discuss American political topics. But because users are permitted to discuss them in a way that offends UK law, there’s the imposition of a UK fine.

The Digital Services Act in the EU is similar. The DSA contains, but doesn’t just contain, content-based regulations and hate-speech regulations. The DSA requires all EU member states to adopt, at minimum, a floor for hate-speech prohibitions. Those prohibitions in the statute are much vaguer than American lawyers are accustomed to.

One of our jurisprudential principles under the American First Amendment is that if you’re going to enact any regulation that comes close to touching speech, it needs to be very clear what you are prohibiting, because you have this chilling-effect concept. A vague prohibition will chill speech, especially when that prohibition is imposed on a large, risk-averse corporation. You impose vague prohibitions on large, risk-averse corporations, and that’s how it becomes illegal to make jokes around the water cooler, for example.

You see the same effect here. The Digital Services Act also regulates other aspects of digital commerce and social media. It regulates things like transparency and competition. We have a lot of Europeans in the audience today, and I hope none of them will find it contentious if I suggest that in Europe there’s more of a focus on technocratic regulation as an arbiter of what’s acceptable than there might be in America, where we have this tradition that really emphasizes rugged individualism and individual conscience.

To be clear, no one is saying—certainly not the State Department or America—“You can’t have your own platforms in Europe. Build your own. Build your own Facebook, your own Instagram, your own Twitter or X, TikTok, whatever you’d like to build, and you can have whatever standards you like on your platforms.”

We’re saying, “These are our platforms. This is our standard, and we don’t want our users or our platforms to be receiving fines. That’s our position.”

I think that’s basically it. When American companies operate abroad, they abide by the laws where they operate. But at a certain point, we recently issued some sanctions, which we’ll get into. One of the individuals we sanctioned was a former EU official who threatened Elon Musk with enforcement action because X, within the United States, had said that it was going to host, on a live Twitter Space, an interview with Donald Trump, our president.

It wasn’t that Donald Trump had said anything violative. It wasn’t that there was a specific piece of content that the EU wanted to ban. It was just that the act of an American business hosting an interview with an American president might offend EU preferences about speech, and that generated a regulatory threat.

When you reach across borders and make a threat like that, it offends American interests and American values. You can expect America to respond.

My history is as an American lawyer in American courts. We’re a nation of 50 states, and each state has its own regulations. We’ve had to think about, when there’s a website in California that operates in Texas, how you decide to what extent Texas gets to regulate.

We have all these jurisdictional concepts. Does the website purposefully avail itself of the forum? Are you posting defamatory statements about a person in Texas? But the mere existence of a website in California that Texas doesn’t like is hardly ever—basically never—a basis for regulation.

When we talk about things like extraterritoriality, what we’re really talking about is this: It’s undisputed that Europeans get to have their own laws in Europe, but we also get to have our own laws in the United States. We’re celebrating 250 years of American independence, and we want our markets to be able to interoperate online. But we’re not willing to give up American freedom of speech in the bargain.

Hey, David, when we look at this—and I’m asking you this one so I can give you a pass on it—what do you think people are so scared of in the UK when it comes to freedom of speech, and maybe the most raucous platform, X, specifically?

I don’t think the people are afraid. I think the government is afraid of the people criticizing it, and therefore they’re engaged in what censors always do, which is protect the people in power.

There’s something, Sarah, you should explain this to us, but as I understand it, there have been over 12,000 people prosecuted—or arrested?

Arrested under the Online Safety Act.

Was that just in 1 year, or is that since—?

That was in 2023 alone. Okay. It isn’t just under the Online Safety Act. I think what’s particularly insidious and particularly relevant about statutes like the OSA and the DSA is that these are portals through which existing censorship laws get applied to the internet.

A lot of these Brits are arrested under existing statutes, like the Communications Act and a law against inciting “racial or religious hatred.” I think we have differences of opinion about what amounts to incitement in America versus the UK. For example, you had a comedian called Graham Linehan who tweeted that if a woman sees a penis in a ladies’ room, she should feel free to kick that guy in the balls. That’s something a lot of comedians say, and I think it channels an impulse that a lot of Americans and Europeans would frankly consider common sense.

But he was dragged out of the airport like a terrorist, had his devices confiscated, was thrown in jail overnight, and lost access, my understanding is, to his heart medication, if I recall correctly.

And because this was an incitement to violence.

Because this offended an existing law against provocative speech in the UK. The Online Safety Act is a device through which all of those existing laws get applied to the internet.

You had another case in the UK where Joey Barton, a footballer, called somebody a “bike nonce.” “Nonce” is not an American term, but I imagine you’re insinuating someone is effeminate for riding a bike so much, or in the manner that he rides the bike. That resulted in a suspended prison sentence, but still a prison sentence.

Because he called somebody—

Yeah, basically. That’s what we call in the United States now. There were some other tweets, too, but none that would meet the bar for American incitement. So, David, you’re absolutely right. That was in a single year: slightly over 12,000 Brits arrested for speech acts. That’s more than were arrested that year in Russia, more than in China, and more than in Turkey.

And when you talk to Brits about this—

You’re absolutely right. Most of the British people that you talk to say this is totally unacceptable. If you look at the polls in the UK, you see public sentiment against this kind of thing.

But I’ve had both public and private engagements with regulators in these countries, and the defense you hear is, “Well, we have a less chilling, less totalitarian environment than China. Maybe more people are willing to break the rules, and more people are willing to offend.”

But if you arrest 12,000 people a year for speech, and you’re raising children in an ecosystem where you can be dragged out of the airport for offending the dogmas of transgender activism, then you might not have a different culture than China for long. Why should the United States be paying to defend your country and support it in fighting, say, a proxy war against Russia if those are basically the values being enforced?

Right, exactly. When we interact with our NATO allies in the NATO context, we hear a lot about our shared history and shared values. It’s time to ask what values we still share.

Together with our allies, we comprised the free world after World War II, and the free world that was assembled against communism. But the cornerstone of a free world, of any free society, has to be freedom of speech. Criticizing uncomfortable speech is where the actual defense is necessary.

Yes. And we have a very special bent in the United States to really go after our leaders. I do it every week with David, since he’s now a public servant. I mean, we go at it.

And you’re knocking on people’s doors strictly for saying, “Hey, I might have disagreements with the Catholic Church. I’m a Catholic.”

Well, a lot of it’s about immigration, right? I mean, I’ve seen a bunch of these examples on X. I saw one clip on X where a judge was handing down a 2-year sentence against somebody—I don’t know if this rings a bell—but supposedly for speech that was criticizing the UK’s open immigration policies. That’s where I sense a lot of the prosecutions are, right?

Right. And this is another place where free speech and freedom of expression are American values and interests in and of themselves. But another priority for the administration is common sense on mass migration, and a lot of the speech that offends those in power has to do with migration policy.

There was a 31-month sentence handed down to a suburban mother named Lucy Connolly in the UK because, after a man called Axel Rudabunga stabbed, I think, a 7-year-old girl, an 8-year-old girl, and a 9-year-old girl at a birthday party, there was ensuing unrest and she tweeted something anti-migration.

It was pretty inflammatory, but it would have been unambiguously legal in the United States. She said, “If this is what migration is going to do to our country”—and I’m paraphrasing slightly, but I remember it pretty well—“if this is what migration means, then burn down the migrant hotels for all I care.”

This was a bereaved mother who’d lost a child. She saw 3 little girls murdered for no reason, and she reacted. Then she felt bad and deleted the tweet. That was a 31-month sentence in the United Kingdom.

Meanwhile, you have actual pedophiles, actual child sex offenders, who get minimal prison time or none in the United Kingdom. That’s led to this epithet that you hear among UK activists: “two-tier policing.” This is the activist cause that they’ve assembled around because they sense that if you oppose mass migration, if you make that kind of critique, you are subject to a different justice system than the kind of person who merely agitates for Sharia law in Britain or merely downloads child pornography in Britain.

Yeah. So, okay, so—

That was the gong of righteousness.

We appreciate it. And then, just to add one more dimension to this: US companies have been getting fined like crazy in the UK and then the EU. I think it’s related to this issue, but can you just describe that?

Yeah.

Because that’s where this crosses over into an ally doing something that we think is mad into directly hurting American interests, I guess, right?

Right. So I don’t believe there have been any big fines under the UK Online Safety Act yet. Its provisions take effect over time, and some of those provisions are just coming online now, including the ones relating to AI.

We have active litigation in American courts right now. One of the leading lawsuits involves the website 4chan, which people who are very online in America may be familiar with. 4chan is a no-holds-barred primordial soup for memes and the like. The cat memes come out of there. Activism—Occupy Wall Street came out of 4chan. But 4chan has essentially no censorship rules. It bans child pornography. That’s pretty much it.

The UK has decided that 4chan is not allowed to exist unless it pays a bunch of money to the United Kingdom for not policing its speech in accord with UK laws—the numerous UK speech statutes that led to prison sentences like the one I just discussed.

There was a large fine handed down during my recent European tour against X. I believe it was €140 million, but that might be dollars.

Are they targeting Elon because they disagree with his influence in the UK?

Look, I can’t speak for the UK regulators, but I can make inferences.

What’s your inference?

X has a particular political valence. We saw Joe Biden, after Elon acquired Twitter, saying, “We’ve got to find ways to go after him,” and I think that sentiment might be shared.

But as an under secretary of state, I’m not an advocate for one American company or even one American viewpoint on the free speech issue. If any American company were fined, let’s say, $140 million by a foreign power for upholding the American First Amendment—if General Motors were treated that way—the US government would have something to say about it.

I also think that X is not the first company to be fined under EU digital regulations. There’s an infographic that circulated recently comparing the revenues raised within the EU through other metrics and then the revenues raised just by fining American tech companies.

Right. And there’s a suspicion that this is really a de facto tax, and pretexts are contrived for fining large American tech companies in order to raise revenue.

Yeah, so that was the thing I think I was referring to, and I think actually the president may have thrown that out. Maybe this is more the EU, but the DSA has become almost like a digital speed trap to try and fine American companies. It massively disproportionately affects them, to the point where you could argue that it’s effectively a tariff on American tech companies operating in Europe.

If that’s the case, Europe is allowed to have tariffs, but then that’s going to change the tariffs that we set. It’s all part of a larger trade negotiation.

Right, exactly. I’ve referred to the DSA before as a censorship tariff because the cost of maintaining the censorship apparatus under the DSA is intentionally levied on specific companies, mostly American ones, that are subject to higher and more intricate regulatory standards than other companies are.

EU regulators say, “Well, that’s not because they’re American; that’s because they’re large.” But the fact that they’re American and not European surely makes them easier as a political proposition to tax.

And so a lot of Americans see this as a tax. Really bizarre, David. We're living in a time where we're seeing freedom of speech and expression go down in Europe and go up in the Middle East. They just had the Riyadh Comedy Festival. There were some rules: You can't criticize the kingdom. Let's leave religion off the table, but you can go after your own. But we might have some sensitivities there.

South Korea does require you to have a Social Security number, essentially, to post online. But, David, I'm wondering what you think about this overall trend in the world of what we're seeing with censorship.

I mean, it's not a good trend. I think that the purpose of censorship, like I mentioned before, is always to protect the people in power. Specifically, it insulates them from criticism. But it'd be a lot better for them to hear that criticism and adjust their policies than it would be to try and switch off the feedback altogether.

It's very clear, I think, in Europe and the UK, that these policies of open migration and mass migration are very unpopular. Why not listen to the people and adjust your policies instead of trying to silence them?

Say what you will about President Trump, and people have varying opinions, but South Park has been deranged this season. They have gone full-bore in attacking him, to a level that I wouldn't feel comfortable explaining the details of here—not on a family podcast.

But even President Trump has a thick skin on these things. We did have one weird thing that occurred. I think it was before your time: the Jimmy Kimmel–Charlie Kirk kerfuffle. But even in that case, it seemed, David, that President Trump and the administration and Brendan Carr, friend of the pod, who's been on a couple of times, kind of rethought that one, yeah?

Well, Jimmy Kimmel was back in the air within—was it, like, 2 or 3 nights? So, yeah, I mean, there's no real censorship there. In that case, it was the network affiliates who were upset because Jimmy Kimmel said something untrue, malicious, and outrageous. So, in any event, the system kind of worked itself out after this problem.

There was no government censorship there. Brendan probably shouldn't have said what he said, in my estimation. Anyway, in any event, there was no government censorship. That's the bottom line.

Yeah, I think it is disturbing that countries that we see as our closest allies, that share similar values and are part of the same Western culture and history, are moving in this direction of more and more censorship. It's disturbing, and I'm glad to see that, under President Trump, the Department of State is pushing back on this.

Thanks, Sarah, for the work that you're doing, and Secretary Rubio. It's extremely important. I think you're making a huge difference. I think we have to use the tools that we have, whether they're tools on trade, the denial of visas, or expressing condemnation, to push back on this as we will and as we can.

Let's talk about some of the new issues. AI—it was pretty obvious 18 months ago, when you saw a deepfake, that it just didn't pass the uncanny valley. Grok images, nano banana from our friends at Google—I mean, these things now, if you're flipping by very quickly, you could make a mistake.

This also, in terms of censorship, raises an issue. We have significant protections in the United States for, say, cartoonists, as do the French, and they're mocking public figures. How is it different when you're mocking public figures—presidents, prime ministers, cabinet members—but the public can't tell because this is new?

Right. I think this is a really interesting question, and it's our privilege to be at this new technological frontier where these new questions arise. I'm glad you mentioned cartoonists. After Charlie Kirk was murdered—and I knew him; I represented him on some First Amendment issues in the United States—I saw Americans walking around in an old T-shirt from 10 years ago. That T-shirt said, “Je suis Charlie,” because that was a T-shirt we bought when free speech in France was under threat, and French people stood up for it.

The Charlie Hebdo cartoonists were bombed. They were murdered for saying things that offended religious zealots. And I think it's a different kind of religious zealotry to not want to allow any dissent online. Thinking back on that episode, and how the European consensus on free speech might have shifted since then, is really sobering. But in America, we take pride in being the kind of civilization where Charlie Kirk and Charlie Hebdo can both speak.

I think making fun of public officials, pointing out when the emperor has no clothes, is one of the most essential things you can do in a democracy. If you believe in self-governance, you have to believe in that. What's interesting about a deepfake is that the point of parody is that you can tell it's parody. But if you're depicting a public official falsely in a way that people can't tell is a satirical or non-authentic depiction, then the parody intention really isn't there.

What I would say, though, is that whenever we reach a new technological frontier, there's a temptation to enact a flurry of new regulations. If we look back over history at other frontiers that have caused similar instability, like the invention of the printing press, people thought that was the end of the world. With the invention of the telegraph, there were worries about disinformation and attention span. With the invention of the film strip, people thought the train was coming at them through the screen.

The impulse to restrain that zeal to regulate, to allow people to adapt, and to give freedom the benefit of the doubt tends to be vindicated over time. When it comes to deepfakes, I think we have, in America and in Europe, strong legal remedies against defamation. So, if someone creates an image of a public figure that is false, and people believe that image—if a reasonable viewer would believe that that person engaged in that action—you can already sue for defamation.

And we have child-protection and underage laws. Those are very strong.

Yes. The thing is, just because you don't have AI-specific laws doesn't mean that you can do whatever you want with AI. You could still use AI as a tool to then break the law and be prosecuted. I mean, if you engage in cyberhacking, for example, and you use AI to do it, you're guilty of a cybercrime.

There are plenty of existing laws that apply to AI, and we should just think about using all of those before you start creating a bunch of AI-specific laws.

Yeah, if those didn't cover it, then there would be a reason to have a thoughtful discussion. I'm trying to think of an edge case here, but for cybersecurity, it's even hard to do. Using voice clones is just fraud, you know?

That's right. Wire fraud, you know? It might be that you use the existing fraud statute, but there are little regulatory tweaks you can make to make the fraud easier to detect.

One approach, which I don't think is always correct but exemplifies one direction of thinking, is maybe there's watermarking or something on some AI images that would mitigate tort liability for some of the providers.

When we invented capital markets on a mass scale, we had our old laws against fraud, but we developed some more fine-grained securities regulations. Now you have to file a certain disclosure annually with your earnings and whatnot. We didn't fundamentally change how we treated false information; we just developed some slightly more fine-tuned devices.

But when I say “fine-tuned,” I think that's an important piece of guidance. You don't just go crazy and try to put the technological innovation back in the bottle, especially when we have foreign-policy rivals like China that are developing AI at an aggressive pace. If we cocoon ourselves in safetyism, we hurt our standing in that race.

So, Sarah, where do you think this relationship between the US and EU is now headed on this topic of free speech? There does seem to be a fundamental divergence. I don't know the conversations that you're having, but do you think this gets worked out, or do you think the divide gets greater? Where's this headed?

So, before President Trump and Secretary Rubio did me the honor of this appointment, I was a litigator, and it was my job to fight. Now I'm a diplomat, so it is my job to be diplomatic. In that spirit, I would like to sound a gong of optimism.

I think that a lot of ordinary Europeans are not comfortable with comedians getting dragged out of the airport, just like Europeans weren't comfortable with comedians getting murdered for publishing offensive cartoons. If you look at polls in Europe, I think you see some of that sentiment.

I don't know where things are going. I can't promise a panacea, but I will say that I've had productive conversations and hope that I'll have more.

If it does become more acute, is Europe prepared for all the American social networks to be turned off and blocked by IP address? Because we really don't need the money. These platforms—it's nice to make money in Europe, but maybe it's time—

Would the European Union, or would various European countries, demand their own version? Would there be a UK-specific version of X? Is that where this would be headed?

To an extent, because one way to resolve the transnational issue is geofencing.

Now, I think I understand that in some of the UK’s enforcement actions, geofencing has not been enough, which is pretty ridiculous. Geofencing is not enough for them. There’s a small American website—I can’t recall the name—that Ofcom has sued, and that website responded, “Well, we’ve geofenced. We’ve blocked UK IPs, so you should have nothing to say about the content on our website.” But Ofcom is still going to—

That should be good, but those people then chose to get a VPN.

Yes, and that’s what—

Which costs $30 a year, and then you can make your own decision. Which is what the people of Iran are doing, too.

You mentioned blocking by IP address. Some countries, where the people don’t have that Charlie Hebdo tradition, like Russia or China, just admit that they are censorious societies and block these websites. It’s like a Great Wall. I mean, the Great Wall in China. We have the Great Firewall, which we’re not trying to take down.

If you wanted to, could that happen? Could the UK put up a Great Firewall and just say that they’re blocking out the outside world? We call your bluff.

I think it’s technologically feasible to a point. There are circumventions, but it is not politically feasible because British people want to be free. I think if Keir Starmer said, “We’re putting up a Great Firewall, and you’re not allowed to access any American social media anymore,” he’d be out of office.

That is their right, but they don’t want to do that because it would be too obvious what they’re doing. Therefore, they want to do the fines. They want to put it underhandedly.

And we’re pushing back, saying, “No, you can’t do that.”

Yeah, good luck doing that against Elon. He’s pretty principled, and I think he can pay for the speeding ticket. I don’t think it’s going to be a problem.

I’d say President Trump’s election definitely changed the direction of free speech in the United States, never mind the rest of the world. Under the Biden administration, we now know from cases like Biden v. Missouri, as well as what was released in the Twitter Files and the further disclosures since then, that the Biden administration was pressuring social networks to engage in censorship.

David, we did discuss this. You could bring up Biden for all of 2025, but when 2026 came around, you had to vote no. I’m just recounting what was happening for several years. President Trump changed that direction. Absolutely. If it weren’t for that, I think we’d still be on a censorship track.

If it was up to Zuckerberg, he would have continued to do it under Kamala. He did it with no problem under Biden. He’s a weather vane. His entire position is based on what makes the system grow.

I think when it comes to these tech companies, there’s a range of courage. Elon is an outlier in terms of being willing to stand up to the government to protect free speech. There are others who just blow with the wind and do whatever the government is suggesting—or demanding. But it was wrong for the government to be doing that, particularly in the US, where you have a First Amendment.

To catch people up, we literally had our FBI putting pressure on our own tech companies, saying, “We don’t like the tone of these tweets, the tone of these posts. We think they’re damaging.”

We would never have gotten to the bottom of COVID and whether we should all be taking mysterious, experimental vaccines. I took it. I’m okay. But the folks who were saying, “Hey, maybe we don’t need this. Maybe we don’t need to give it to kids”—that whole discussion was shut down by the Biden administration.

Irrefutably. The pretext for some of this was disinformation, a term that was really distended to encompass anything. If you read the white papers put out by these disinformation NGOs, they will admit, “Yeah, the information can be true, but if it promotes an adverse narrative, we don’t like it.” That’s such an Orwellian adverse narrative. They call it misinformation.

No, not misinformation. There’s misinformation and disinformation. The way some people define these is that misinformation is false and disinformation is bad. If disinformation pollutes your democracy, the wrong candidate might win, and I think that’s really the impulse.

But we had information suppressed under the auspices of combating disinformation that turned out to be true. Things that were suppressed included the assertion that the vaccine did not completely prevent transmission. That turned out to be true. It mitigated transmission significantly, but it was not a sterilizing vaccine.

The reason why a lot of people took it is that they didn’t want to be a blocker in the system. It was, “It’s my social obligation to do it.” If they had been given the choice, they might not have.

Another thing that was suppressed was the assessment that the virus might have leaked from a lab. We now know that was the same assessment of a House committee and the CIA. So the government thinks the virus might have leaked from a lab, more likely than not.

You’re reaching out to Twitter, Facebook, and Instagram, saying, “Hey, we can’t force you to take these posts down under the First Amendment, but we’d really appreciate it if you did. And if you want to stay in our good graces, you should.” Which is a long way of saying that we all need to be vigilant about it. In the United States and internationally, if you’re not vigilant about free speech, there are people who will take it away.

Yes. Let me ask a question about that. You mentioned organizations or NGOs that are instigating regulators. They’re ginning up these regulators and showing them cases: “What about this? What about this? What about that?”

I’m curious because I think you’ve called this a censorship-industrial complex. Could you explain what this thing is? Do you think that some of these groups, which are in the US and not just Europe—in fact, they might mostly be in the US—are going to European regulators as an end run around the First Amendment? They can get European regulators to censor material that otherwise cannot be censored in the US.

It’s a great question, but it’s a question we don’t even need to ask because we know the answer, and the answer is yes.

We have emails that have leaked from some of these NGOs. One of them, the Center for Countering Digital Hate, is a British NGO whose leader was the target of some of our visa sanctions. There are emails exchanged with Democratic politicians in the United States and with politicians now very close to Keir Starmer, saying, “Our number-one priority should be to kill Musk’s Twitter.”

So, kill an American company in order to suppress American political speech, and our second priority is to instigate UK and EU regulatory action. This is an entity taking government money to get foreign governments to come after American businesses.

This whole fact pattern, where these American NGOs were working with the American government to send forceful—but, they allege, not technically coercive—emails to Twitter and Meta, was an attempt by these activists to replicate the EU DSA in a way that would dodge the American First Amendment.

The EU DSA requires member states to designate NGOs as so-called trusted flaggers. This organization’s job is to sit on Twitter, look for offending tweets that might be hateful or whatever, and report them to Twitter. They get a privileged reporting channel, and the companies are required to give those reports first-tranche priority.

If you look at what was happening under Biden, it was a very similar system. These government agencies would arrange first-tranche priority for these reports, and some of what was put into those channels—the reports—were technically made by NGOs.

You also had one of the small upsides of COVID, if you care about government transparency: everyone was holding their meetings on video. So you have these videos of Zoom meetings with government operatives saying, “We couldn’t do this under the First Amendment, but fortunately this NGO on this call with us is going to do it instead.”

They have ways of pressuring people, which leads to our final questions for you. There are firms that started in the conservative space and moved to the liberal side, trying to get advertisers to cancel on this program. They went after Howard Stern, and they went after the liberals first and then the conservatives. “Let’s get Rush Limbaugh’s advertisers to cancel. Oh my God, he said these incendiary things,” yada yada.

But we had an even more pernicious one, which is when people started to say, “Well, hey, you’re Cloudflare. Hey, you’re Amazon. Hey, you’re PayPal. Hey, you’re Stripe. We’re going to go after you and make sure that we shame and pressure you, sometimes behind the scenes, to debank and demonetize.”

YouTube got pulled into this as well. “We’re going to shadow-ban your videos.” They started labeling All-In videos because we had conversations with scientists about COVID.

Okay. Labeling also suppressed, I think. What are your thoughts on that? It seems even more pernicious because if you take away a person's ability to monetize it, how do they scale back?

So, there's been 1 successful Supreme Court case in U.S. history on viewpoint-based debanking, and that was my case, which we won. That was called NRA v. Vulo. The way we got that into court was—you had the New York financial regulator, for the pleadings, literally reaching out to financial institutions saying, “You know, it would really be better for your enterprise risk-management framework if you didn't do business with any pro-gun groups.”

Enterprise risk-management framework.

They have all these professionals, and you guys have been in finance, and you know this. These bureaucracies are hired within financial institutions to ensure compliance with all these regulations. They have these elaborate risk-management protocols. This gets a bit into the weeds, but there's this thing in finance called reputational risk. That's supposed to be the reputation of a bank for safety, solvency, and soundness.

You don't want to run on the bank. If everyone thinks the bank might fail, that's bad for the system. But there's this ESG movement to expand the concept of reputational risk to include things like, “Do you have a reputation for letting naughty, disfavored speakers have bank accounts?” And that came up in the NRA case. The Supreme Court says the government is not allowed to do that.

Even though the government—our First Amendment says, “Congress shall make no law” restricting the freedom of speech—this wasn't Congress making a law restricting the freedom of speech, but it was a government entity adversely applying regulations to choke off certain viewpoints. Instead of going directly to the guy saying the thing you don't like, they were putting pressure on this risk-averse middleman, this bank. The debanking and deplatforming is insidious for exactly that reason. When you have a risk-averse middleman like a financial institution, it's almost like they designed it that way.

Yeah. They don't have skin in the game with respect to it. They don't believe in your speech the way you do. They have their in-house counsel telling them that this is going to piss off the financial regulators, so it's easier to take it down.

The common theme is that when the government can't do it directly because it would be a violation of the First Amendment, they use an intermediary to do it.

Right. So, you get the bank to debank someone, or you get a dark NGO.

Which is really a government organization that's funded by the government, but they call themselves nongovernmental. So, they do the quote-unquote fact-checking.

Yeah. Or you get one of these other cases where the FBI, through the Biden administration, is putting pressure on the social networks. In any event, you get, like you said, a middleman to do the dirty work because the government can't do it directly.

And to do it in a really nefarious way that's hard to detect. It's like, “It'd be a shame if we blocked a merger.” Zuckerberg's a pragmatist. I'm going to go at him again. He likes to buy things, and if the FBI's calling you and you've got to get something through the FTC next, you're going to try to make nice, right?

That is the risk of giving any regulator a capricious cudgel over the internet. Even if the regulation isn't explicitly speech-based, it's just, “You can only do your merger if this guy likes the look of the merger,” then companies are going to vie to impress that regulator. I think a bit of that was going on, frankly, with the Jimmy Kimmel thing.

You had this merger in the works, this Techna merger that they tried to complete under the Biden administration. My friends in telecom tell me that when Techna was trying to sell itself during the Biden administration, it went out of its way to show the regulator how woke it was. So now it has an incentive to show the administration that it's MAGA-aligned. That's what happens when you give a regulator a large cudgel.

Now, I want to say something about labeling. Labeling videos sounds like, “Oh, it's just transparency. What could be wrong with that?” People should know if fact-checkers think that something is wrong. I think my office's approach on that is that it depends on who's putting the label on there and for what purpose.

So, a lot of these disinformation NGOs—it was almost like McCarthy, almost like the Red Scare.

Yes.

They'd make a list of outlets that were spreading disinformation, but they wouldn't just publish that list. They would send it around to the credit-card companies and payment processors and suggest, with an implicit government imprimatur, because they're all government-funded, as David very relevantly points out, “You guys really shouldn't be funding these websites.” The websites would never know why, and the viewer would never know why.

I think a type of labeling that's really good is the type exemplified by Community Notes on X, where I can read the tweet, and then I can see what the Community Notes say about it.

You can see it all. You can see a ranking of the Community Notes. It's very interesting.

It's been a total game-changer. I remember when they were doing fact-checking, and the fact-checking was so bad because the fact-checkers were biased, or sometimes it just wasn't good quality control. But the Community Notes thing has really worked, and you'll see that when someone posts something that's truly misinformation, like a fake image or a fake article or something—

Like—

It always gets caught. It's kind of timely, and then you get notified. You get a notification. Have you noticed this? If you like something, it'll circle back.

It gets Community Noted. I get a notification, and then I feel like an idiot: “Oh, I got fooled by this.” But people will drag social media for the fakes or whatever, but I never get notified when The New York Times makes a mistake and posts a correction on page 43. They never notify anybody about that. They bury it on the last page. So, I still think social media is by far the best.

Were you in the meeting when they were deciding Community Notes, and Elon was like, “Tell me about it”? And I was like, “Elon, I think this is really interesting. You should—”

Yeah. Double-click on it because it's actually working. He looked at it, immediately understood the algorithm, and said, “Keep the group.” That group stays.

Right. They look for consensus among rivals, which is fascinating. I'll tell you, the other game-changer on X has been Grok, because you can just go at Grok and ask, “What's the truth?” You won't necessarily always agree with Grok. I'm not saying it's perfect, but it's pretty darn good.

It's trending in the right direction.

It's—I mean, it's really good on the whole. It does a really good job fact-checking, too. So, we don't really need these bureaucrats, politicians, and regulators telling us what's true or not. We have Community Notes, we have AI now, and you've got other users—

—who can file a lawsuit if you feel you've been defamed. This exists in the United States as a concept.

Listen, Sarah, I think I can speak for everybody here in USA House: We're really glad that you're so vigilant and dogged in protecting the First Amendment. Give it up for Sarah Rogers.