Evelyn Douek:

Social media platforms do provide this avenue for government jawboning, so the setup being this is a concern to take seriously. Now the question is what form does that concern take and how do we deal with that problem?

Hello and welcome to Moderated Content, podcast content about content moderation moderated by me, Evelyn Douek. We’re coming to you midweek this week for one of our highly irregular long form interview shows to talk about an order out of a US district court this week in joining large swats of the government from talking to platforms about content moderation in almost any way. Judge Terry Dowty in the Western district of Louisiana issued the order after finding on Tuesday that a number of plaintiffs, including the states of Missouri and Louisiana were likely to succeed in establishing that the government had improperly pressured social media platforms to remove content, including about COVID-19, the Hunter Biden laptop story and the integrity of the 2020 US election in violation of the First Amendment.

Now we’re going to try and do something very difficult here In this episode. We’re going to try and have a nuanced conversation about genuinely hard legal and policy questions. The politicized and polarized climate around this topic in particular has not been conducive to that kind of conversation, and ironically, given the fact that this debate is notionally about the importance of free speech, the result has been that many people are chilled from speaking about the complexities here. To avoid making this conversation all about the politics, we’re going to try and place this district court opinion in its broader legal context, so be warned, this conversation is going to be very doctrinal and very wonky. It will not have hot takes and it will not take petty pot shots like pointing out, for example, that the opinion starts with an epigraph, that misattribute a quote to Evelyn Beatrice Hill rather than Evelyn Beatrice Hall, where we’re above that kind of thing here. There are many other places where you can go to read or listen to that kind of content.

This is going to be a serious conversation about what is legitimately a very serious and difficult issue. When I want to have serious and nuanced conversations about the complexities of First Amendment doctrine, there’s one person who always delivers, Genevieve Lakier, professor of law, and Herbert and Marjorie free teaching scholar at the University of Chicago. Genevieve has written and is writing about the First Amendment law of jawboning, and so she’s the perfect person to talk to about this. Thank you Genevieve for agreeing to share. You’re extremely cool, but not chilled takes about all of this.

Genevieve Lakier:

Well, thanks for having me, but now I’ve got a lot of pressure to be cool, but not overly chilled.

Evelyn Douek:

That’s right. Get the temperature just right. Not too hot, not too cold.

Genevieve Lakier:

I’m not typically described as cool, but I will do my best.

Evelyn Douek:

Excellent. Okay, so let’s start at a very high level because I want to situate ourselves in this broader conversation. This order is about something that legal scholars will often talk about as jawboning. What is Jawboning and what does the First Amendment have to say about it, generally speaking?

Genevieve Lakier:

Okay, well, jawboning is a term that has quite a long history. It goes back to 50, 60 years, goes back to the early 20th century, and it is generally defined or used to mean the use of informal methods by government actors to pressure private companies into doing something that the government officials cannot actually require the company to do. You are using informal methods to expand as a government official your authority, but today it’s most commonly used to talk about efforts by government officials to pressure private speech platforms or private speech intermediaries like the social media companies into suppressing what is otherwise thought to be protected, constitutionally protected expression.

Evelyn Douek:

Just so I’m clear, jawboning can mean any kind of intermediary, but today we most commonly associate it with speech intermediaries. Is that what you’re saying? And does it matter? Does that distinction matter? Is there something that we should be particularly concerned about if it’s speech intermediaries?

Genevieve Lakier:

Yeah. I think historically the first time the term was used, it was to talk about efforts, I think by the Kennedy administration to make steel companies limit their wage and price increases, so it had nothing to do with speech at all, and it’s been applied in the banking context. It is applied generally when we’re thinking about government pressure on private companies, but when it involves speech, this both raise a special constitutional concern because of course, according to the First Amendment, speech is special, and so government efforts to regulate other goods and services in the economy just have a different constitutional status than government efforts to regulate speech. We might think that it’s particularly worrisome when we’re talking about government jawboning of speech intermediaries because it’s well recognized that these intermediaries that there’s good reason to worry that they are going to not fight back, not push back against the government pressure too much, and this is because of their incentives.

Think about, say the Kennedy administration trying to get a steel company to limit its prices. The steel company is going to have very strong incentives, economic incentives to keep those prices up because that’s how it makes profit, but a large speech intermediary, say a social media platform like Facebook or Twitter, it’s revenue comes from advertising and also from just being a site where lots and lots of speech flows through the platform all the time, meaning that there’s a big audience, meaning that advertisers get a lot of eyeballs. It doesn’t particularly care about any particular speech act post or tweet, usually at least. And so the government comes to it and says, “Hey, we want you to take down this particular tweet or post.” And this is often going to be speech that is controversial, maybe advertisers aren’t going to like it. We might worry that the speech intermediary is not going to have a very strong incentive to fight back.

And so this is jawboning when it comes to speech intermediaries, we might worry not only affects constitutionally protected interests, interests that are really important to democracy like the diversity and vitality of our public sphere, but also are more likely to succeed than it might be the case with respect to other businesses. I think it makes sense that when we talk about jawboning today, the primary thing we’re worried about is jawboning of speech platforms.

Evelyn Douek:

Right, so I think that’s extremely helpful to underline the particular role of the intermediary here in saying that they don’t necessarily have adequate incentives to push back. They don’t necessarily care about a particular post of an individual. There’s also other advantages maybe for the government in going to pressure a social media platform as opposed to trying to go and find every single individual user that’s tweeting a certain kind of contents. There’s also transparency concerns around if you’re pressuring the intermediary and not nec… You can do that in a way that might be more opaque than going and pressuring individuals and the effectiveness of doing that through the intermediary. There is something special about using the intermediary that raises all these particular concerns and is being raised in this context, and so a very leading question, but is the First Amendment cool with this? How does the First Amendment think about this generally speaking?

Genevieve Lakier:

Well, and just to add on to what you said, we might think that there’s all kinds of problems when the government uses this pressure even to go after unprotected speech or expression stuff that has been considered to be outside the scope of the First Amendment, just because ordinarily what we think a good safeguard against bad government behavior is political payback. If the government passes a law that is overly repressive of things that we care about, and even if the Constitution doesn’t protect it, we might think that the ordinary democratic process is going to kick into action. At least theoretically there are problems with American democracy, but this is at a high level of generality, there’s going to be political pressure on whoever’s passing this bad law.

But when it comes to things like jawboning, so much of this happens in the shadows. There’s private meetings, there’s private phone calls, there’s emails no one has access to that is very hard for the ordinary democratic process to operate, but all of that stuff that’s true in general of all kinds of informal off the scenes government behavior and the First Amendment doesn’t have anything particular to say about this kind of non-transparency. It’s when the government tries to use these informal methods to skirt the boundaries of the First Amendment that courts start to get interested or worried.

The general rule, in theory at least, we’ll talk about how maybe it doesn’t work so well in practice, but in theory at least is as the court said in a, maybe it’s most famous jawboning case, Bantam Books v. Sullivan, that the government cannot do informally what it cannot do formally essentially. If it is using pressure to try and get a private speech intermediary to suppress or take down speech that is protected and that the government cannot directly ban itself, that violates the First Amendment, even if it’s done by means of threats and letters, not by means of laws or agency rules. That’s the theory, in practice though making sure that’s actually what’s going on is a difficult question, but for seven years, that has been the doctrinal rule.

Evelyn Douek:

All of that is background, doctrinal background, theoretical background just to say this is a legitimate concern that something that the First Amendment has been concerned about and something that scholars and academics and researchers have been concerned about for many years and decades. Even in this context as well, the social media context that social media platforms do provide this avenue for government jawboning. The setup being this is a concern to take seriously. Now the question is what form does that concern take and how do we deal with that problem? So let’s then enter the conversation about this district court judgment that came down. On independent day July 4th the decision is 155 pages and the recitation of the relevant facts goes until about page 86, and so asking you to summarize the facts is definitely unfair, but I’m going to do it anyway.

I do think though that part of the problem with this conversation and the difficulty of having public conversation is that there are so many facts, and the opinion doesn’t do a great job of distinguishing between different kinds of contact between platforms and government actors and lumps them all together in one big giant bucket. And so you’re left with this impressionistic milieu of what’s going on rather than being really specific about what’s going on and that allows people to reach their preferred conclusion. Can you talk a little bit about what exactly was the concern in this case and what are the kinds of contacts between government officials and social media platforms that this judgment talks about?

Genevieve Lakier:

This opinion is quite, it’s overwhelming even for someone who really loves to read First Amendment opinions of all sorts, because this suit is so ambitious and clearly politically motivated. It is clear that all of this is happening both because of I think growing concern about the problem of jawboning, but also part of a effort by, in this case attorneys general from Missouri and Louisiana to find the Biden administration guilty of what they call the biggest campaign of censorship or speech suppression in American history, which there’s just no way. There’s been, as a First Amendment person, there’s been so much that’s a very high bar, so we’re nowhere knowing close to that.

It is a bit over overwhelming, but in stepping back, thinking about just where we are on the case and what happened, so in August 2022, there was a lawsuit filed by Missouri and Louisiana Attorneys General and then a number of private plaintiffs, some of them COVID vaccine skeptics, some of them I think the founder and operator of the Gateway Pundit, the far right online news publication who claimed that their speech was being suppressed on Twitter and Facebook and other large platforms as a result of this conspiracy of government pressure and inducement, and it’s just an amazingly ambitious lawsuit. It cites as defendants President Biden, the Surgeon General, the press secretary, the Director of Health and Human Services, Anthony Fauci, and then a large number of government agencies, all of whom it claims were involved in suppressing speech or pressuring the platforms to suppress speech about the origins of COVID, about the Hunter Biden laptop, about vaccines and mask mandates, about voting fraud, about a huge array of issues.

And then the evidence it provides… Oh, and the important thing also to note is that usually, for reasons maybe we can talk about in a little bit, usually when you get allegations like these in federal courts, it’s very, very rare. In fact, I don’t know if any other case besides this one for the plaintiffs in these cases to get to discovery, which is when courts allow there to be disclosure of emails and other private information to try and substantiate the allegations in a complaint, but in this case, the district court Judge Dowdy, who’s clearly very sympathetic to the plaintiff’s, claims, he allows there to be limited discovery for purposes of a preliminary injunction motion, and so there’s depositions of Anthony Fauci and other government officials. There’s a lot of additional information added to the record.

And then what came down on July 4th was the after all this discovery, the order and the opinion on the preliminary injunction, and as I’m sure all the listeners know, this rarely remarkable, very broad opinion agreeing with the plaintiffs on basically all of their allegations that there is a likelihood that this menon was violated and therefore the huge swaths of the federal executive branch can be enjoined from speaking to the platforms about speech on their platforms. In order to get there, just as you said, the court just picks up and articulate, resuscitates this wide array of “damning” facts, damning in quotes, that the plaintiffs had put in their complaint ranging from things that on my view are not worrisome at all, to things that are more worrisome without distinguishing between them at all. The range of things that are identified are in some cases just completely public statements like Biden’s infamous statement that the platforms are killing people.

Two instances in which the platforms come to a government agency to ask, for example, in one case they ask, they go to the CDC, I think to ask whether a post by Fauci actually was by Anthony Fauci or was by an impersonator of Anthony Fauci, which is information of course that you would understand why you would go to the CDC to find out the platform may not know, but the government agency presumably would. Two things that are more worrisome, there are these few emails by people in the White House who send emails. In one case it’s to Twitter about a parody account that is about, I think President Biden’s granddaughter and in which they ask this, Rob Flaherty sent the email and asks. I think it’s a command, I don’t know how to read the email, but says something like, “Cannot stress the degree to which this needs to be resolved immediately. Please remove this account immediately.” And Twitter does.

That feels like an incredibly different kind of government action to the CDC responding to a platform’s request to identify whether a particular post is or is not from Anthony Fauci, but this is all swirled together into a soup to make this, to reach the conclusion that there’s this vast conspiracy and the district court does not distinguish between them at all.

Evelyn Douek:

Yeah, it is incredible though the vast array. At some point it’s talking about the fact that Dr. Fauci had this exchange with Mark Zuckerberg where they ultimately set up the fact that he did three live streams on Facebook with Q&As about COVID-19 and made a public announcement, and it’s all this kicking up dust to try and say, “Look, these government actors and these platforms are intertwined and they’re so closely working together in how they think about these things in a way that isn’t particularized in the normal way that we think about these things.”

Two, I agree with you, and to be clear, this is something that we’ve both talked about and written about before, much more concerning things. I think there are legitimate sort of speech interests and speech values that we should be concerned about when there are these individual emails about individual pieces of content or individual accounts with very aggressive language from the government to platforms saying this needs to be removed. I think that’s something that I think both sides of this tend to be too blythe in sweeping away the concerns of the other side in not recognizing the problems here one way or the other. The opinions, lack of discrimination between those very different categories of behavior, I think makes it especially difficult to wrestle with.

Genevieve Lakier:

Well, I find it frustrating because for a number of reasons. One, I worry that the really problematic stuff is going to be lost in the smoke or what did you call it? The dust, that there’s so much going on that it is hard to identify what is really problematic. For very understandable reasons, I think it is going to push some people to dismiss the opinion as just terrible, just politicized judging, which clearly this judge is not working very hard to do a careful job of thinking about or working hard at all to think about the trade-offs between, and we can talk more about those, between the free speech rights of the users and this desire to not have the government manipulate in the public sphere versus on the other hand, these really significant interests in having there be forms of information sharing between government officials and platforms and also, ultimately some kind of democratic accountability or democratic oversight over these incredibly powerful private speech regulators.

There are speech interests on both sides, and the district court doesn’t try at all to do the difficult work that is required of it here, I think, which is to think about how these play out to balance them and to reconcile them. That said though, so it’s really frustrating. I understand why people will reject it, think it’s a terrible opinion and should just be tossed in the garbage. On the other hand, there are the Israeli concerning emails, and also I’ll say this lawsuit’s been going on for a while and it has been clear from the beginning where it’s heading. Although maybe the breadth of the injunction still surprised people, certainly surprised me, but it’s been clear that the judge is sympathetic to the plaintiff’s case and that it’s likely that they’re going to do well in this lawsuit, and that has led to a lot of concern and worry about the politicization of the courts understandably.

But on the other hand, I have to say that I am grateful that there was discovery in this case, because up until now there’s been very limited information about what exactly is going on, that we know that there’s lots of contact between government officials and the speech platforms and officials at the various social media companies, but it’s hard to know what those consist of, how often they are, how troubling they are. You just don’t know what you don’t know. Although there are so much, there’s so many facts, we still don’t know clearly the deposition. The discovery was limited and this is only one case, and of course it was also designed to reach the conclusions that the plaintiffs wanted, and so there are other facts that were not discovered.

It’s very interesting to me both what we know and also just reading these emails, the tone that they take, what it suggests about the way in which government officials and people in the platforms were understanding the nature of their relationship. For example, in some of these emails, quite a few of these emails, the government officials doesn’t say I or you, he says we, “Can get this taken down? Should we do this?” That’s really interesting. It may just be performative a way of trying to get the platforms on board, but it does suggest a familiarity or joint action. That raises a lot of questions for me as a First Amendment person who believes that there needs to be some kind of separation between the public and the private when it comes to speech regulation. Then here too in this email, the assertiveness of the email, “This needs to be taken care of immediately.” Was interesting to me and it makes me curious about the nature of these communications.

Evelyn Douek:

I would definitely agree with you about the value of transparency. I think transparency around this is something that people have been calling for a very long time. We can have a conversation about what kinds of contact is appropriate and what is not, but we’re in a situation where often we don’t even know what kinds of contact are occurring, and so bringing transparency to those processes can be valuable and I think it allows the democratic processes to operate. If one of the main concerns we have about jawboning is it’s end run around democratic processes, then transparency feels like a good fix for that.

This judgment of course calls into doubt, perhaps the or raises limits around using discovery or the judicial system as it currently is constituted as a mechanism for that transparency, because as you’re saying, the facts are overwhelming and it’s clearly presenting facts in a way that’s intended to paint one picture and is not a comprehensive accounting of the facts. It pulls quotes out of context. It does things like say government officials requested this particular account to be removed. It was removed and doesn’t note, for example, that there are several months going by between those two things happening.

And so it’s clearly painting a particular picture that’s not the kind of transparency that’s necessarily helpful for democratic processes to work. In general, the idea that all of this has been going on, we don’t know very much about it, and courts generally have been dismissing it, I think, and that’s a problem I think is right.

Genevieve Lakier:

There’s an irony here in that the opinion suggests both that why we should care about the problem of jawboning, that there are real free speech interests at stake here, and also how dangerous it is to give courts broad discretion to find that the government has violated the first amendment via jawboning. I think it raises lots of questions about the role that courts are supposed to play in safeguarding and patrolling free speech values. There’s this broader question, maybe we can say that for another wonky podcast about the extent to which we want this to be a First Amendment question. Do we want this to be a regular? If we are going to protect against jawboning, if we recognize that there’s some forms of government platform relationships that are concerning, maybe it should be legislation, maybe it should be regulation rather than judicial opinion. In the absence of anything else, what we have right now is this what looks like a struggle going on in the courts, and that’s got its benefits and as you say, definite downsides.

Evelyn Douek:

Right, so I want to get to that struggle and the doctrine soon, but just before we do a couple of other things to note. One of the things that the facts don’t make clear but should be clear is that this pressure didn’t always work and the judgment is literally citing times where there were these relationships between platforms and government actors, but it wasn’t that the government was just requesting things and the platforms always acquiesced. There were many times where platforms would respond politely, “No, this one doesn’t fall outside our rules, this one does. We will take it down.” I think the judgment itself ironically makes it clear the kind of independent review that is going on within the platforms, and it’s not just that they’re acting at the behest of the government. We’re going to come back in a second to ask how important that is.

But before we do, I just want to go to the issue of threats and talk about what were the threats in this particular case that the judgment talks about, because I think when we come to the legal standard and talking about what are the kinds of problems that the First Amendment is concerned about. Understanding it’s not just that the government was having contact with the platforms, but there’s some idea that the government was threatening the platforms, and so what were the threats that the opinion lays out in this case?

Genevieve Lakier:

I think the opinion talks about two primary threats and certainly the complainant did, so one is the threat of limiting Section 230 immunity, which is often raised as a threat that is thrown around as a basis for jawboning claims and that’s because members of Congress and Biden himself has suggested that 230 immunity should be limited because the platforms are doing a bad job regulating speech that flows through their platforms. And so obviously would affect the platform’s bottom line, so I think that’s the primary threat. Then there’s also talk about antitrust regulation or litigation to make the platforms more susceptible to antitrust claims, which also affects the platform’s bottom line. These are slightly different kinds of threats because I do think when we’re thinking theoretically or normatively about what’s the kind of threat that violates the First amendment, what’s the kind of threat that doesn’t.

It does feel to me like it might be to say, if you are doing a bad job with this kind of speech, we may pause regulation to make sure you do a better job. That feels to me more legitimate then, “We don’t like how you’re dealing with the speech, so we’re going to use this completely separate regulatory tool to punish you. It has nothing to do with the speech. It doesn’t reveal our authority to regulate the speech directly.” And so the antitrust threat actually to me seems somewhat different from the 230 claim, because the 230 claim might just be saying, “We think if you’re going to do a bad job, if you’re going to be irresponsible about speech, then you should pay whatever financial penalties the First Amendment allows you to pay for harms that come from that speech.”

I think implicitly if you read through the facts, there’s another threat that is being wielded and that you often see in jawboning cases, which is a threat of bad publicity that government officials are going to use the bully pulpit to criticize or impugn the reputation of the companies if they don’t do what they say. I think this is often understood to be a pretty credible threat, but it’s a different kind. It feels more like a democratically legitimate threat, but I think it is one of the things that was going on.

Evelyn Douek:

Let’s get wonky and doctrinal then as we promised and talk about the legal test, and what is the test that the First Amendment applies to say this kind of contact between a platform and a government actor or an intermediary and a government actor is a problem and what tested the court use in this particular case?

Genevieve Lakier:

Okay, yes. I do want to convey to the listeners how incredible this opinion is just as a pure work of doctrine.

Evelyn Douek:

Incredible in a literal way?

Genevieve Lakier:

It’s hard to believe, it’s really a leap into the unknown. In some ways one of the things that I think incredible about the opinion is that it recites the conventional test. Okay, so some background. As I said earlier in the 1960s in a case called Bantam Books v. Sullivan, the court says, recognizes that these kinds of informal pressure tactics can in fact violate the First Amendment and that there are state action in these cases and suggests that something like the test, this is an opinion about Justice Brennan. It’s during the Warren Court, so it’s a very functionalist realist court. It’s not very formalist, so it’s not a very tightly articulated test, but something like if the recipient of government pressure would perceive this to be coercive, that they have little choice but to comply, otherwise they’re going to face sanctions. If they would take it as a command, not a request, and then they do take down the speech, then that violates the First Amendment, and that’s the test.

It’s a totality of the circumstances, analysis, we look at the facts and the circumstances and we try and put ourselves in the position of the recipient of the threat and say, “Is this coercion or is this an attempted persuasion?” And if it’s coercion, it violates the First Amendment. By the 1970s, however, we have, the court’s changed a lot and it’s much more formalist, and although this is now the Burger Court, not the Warren Court, still pretty committed to the First Amendment, but has quite a different vision of the First Amendment. It decides a variety, a number of really important state action cases, and one of which is this case called Bloom v. Uretsky. Not the most memorable name, and the facts are quite complicated and not that sexy. It’s about whether or not regulations that govern private, but Medicaid… Private healthcare facilities that receive Medicaid funding, but that the regulations encourage these healthcare facilities to either give less intensive care to some patients or to stop hosting patients. This is all designed to save taxpayer money, so we’re trying to encourage these clinics to be less generous with taxpayer funded Medicaid dollars.

Whether that those new regulations are enough to establish state action in a case where there’s this due process challenge, people are kicked off their healthcare or they’re excluded from the clinic and they don’t get a hearing and they want to bring a due process claim. Not a First Amendment case, not a jawboning case, nothing to do with the specific First Amendment issues at stake in these cases, but the court says, speaks generally and says there can be no state action except when the government either, I think the language is coerces or significantly encourages private action to such a degree that the action can be fairly said to be the state’s own, it has responsibility for the action.

Basically it’s only when, we can realistically say it’s the government acting, not the private party acting that there is state action when a third-party is harmed. Note that this is a really different test than the one that the court was talking about in Bantam Books. It’s not about would the private party, in this case, the agency, think that it’s being coerced or think that it has to act. It’s some kind of objective judgment about whether the coercion was so strong that essentially the private party had no discretion, had no choice but to comply with what the government is requiring it to do.

Now, there is no requirement that this standard, which is about due process and about state action generally had to be read into jawboning cases. The most on point case is Bantam Books, which as I said, has this somewhat loosey-goosey, but much broader notion of when this jawboning or informal pressure could satisfy state action and violate the First amendment, but when we start to see social media jawboning claims hit the courts in the 2000s and 2010s, and they’ve been intensifying in numbers, there’s just been lots and lots more of them over the last, say, five years. Courts have relied on this Bloom v. Uretsky test or language to develop a very, very narrow test of when the First Amendment is violated by informal government pressure on social media platforms to kick off speech, and basically have held that there has to be some kind of direct government involvement or direction of the particular, the decision to suppress or some very, very serious threat. That just saying there was pressure and maybe even the pressure worked is not enough.

That’s the test. That’s the test that the… And that’s the precedent against which the district court is operating. What’s so interesting about what the District court did, what’s so incredible is that it takes this test and it doesn’t fight it. It doesn’t say, “Oh, I’m rejecting Bloom.” This is I’ve suggested that maybe court should reject Bloom and they should use Bantam Books, which is more on point and they should say, “We’re just going back to Warren Court and we’re doing a socially realist notion of how people perceive government pressure, blah, blah, blah, blah, blah.” That’s not at all what the district court does.

It says, “No, no, no, Bloom’s the case and I’m going to rely on it.” But there’s this language in Bloom about significant encouragement. Does the government coerce or significantly encourage the private party to such a degree that the action can fairly be held to be the responsibility of the state? It ignores that to such degree that the action can fairly be held to be the responsibility of the state, and it basically ignores all of the prior precedent that says, “Bloom’s a really, really, really narrow test and there has to be basically no private discretion whatsoever for Bloom to be satisfied.” And it just says, “What does significant encouragement mean?” Well, it means that the government’s encouraging. I think what it takes significant to mean is just on multiple occasions, if you encourage and then you do it again and you do it in enough times, that’s significant encouragement.

It’s not looking at whether the platforms themselves felt like they were coerced. Maybe they did, but that’s not what it’s looking at. It’s not a Bantam Books kind of analysis, and it’s not asking, is there no discretion? Because clearly there was in, as you said, in many of these cases, the platforms just reject the government’s requests. It just says, “Were they encouraging and did they do it a few times?” And finds that’s enough to violate the First Amendment, so it’s a totally new test even though it purports to not be.

Evelyn Douek:

Right. It takes these words out of context and just looks at, “Well, here’s the definition of significant, here’s the definition of encouragement, and we’re going to use those without looking at the cases that it came from.” And it totally rejects the idea that how the platforms felt about it really matters at all. Indeed, there’s this passage where the judge’s rejecting the argument that the platforms are raising or that the lawyers are raising to say, “These platforms would’ve done it anyway. The pieces of content that they took down were violations of their own rules. These were pieces of content that they have declared they don’t want on their websites, and they would’ve taken them down anyway.”

And so the government action here it, there’s no but for causation necessarily. It’s not that this is stuff that they wanted to take down, and the court rejects that and says, “Look, it doesn’t matter if the platforms would’ve taken it down anyway. It makes no difference what decision the social media companies would’ve made independently of government involvement, whether there’s evidence that demonstrates the wild scale involvement here.” So the company impression of whether they were being coerced or whether they were being encouraged or whether they even wanted this kind of engagement, which we will talk about, I’m sure is irrelevant. How do we think about that? The complete dismissal that the companies would’ve done this anyway, so it doesn’t matter, how should we think about that?

Genevieve Lakier:

I have somewhat mixed feelings about that aspect of the opinion because if used incorrectly, that could be a get out of jail free card too easily to say, well, the defense that is often made in jawboning cases is, and I think explicit strategy on the part of government actors is to say, “Well, we’re not saying that the platform should take down speech because we think it’s infringing. We’re only identifying speech that is infringing under the platform’s own rules.” That’s true, but if the government is bringing to the attention of the platform only say democratic speech that happens to violate the rules, and the result is that 90% of that speech is taken down and similarly situated Republican speech is not taken down, and that’s a result of the government’s intervention, I think we might worry that’s still manipulation of the public sphere and raises free speech concerns. Even if in some theoretical way the platform could say, “We would’ve done it anyway because it violates our rules.”

And it’s also true in jawboning cases that even in cases where the court finds that there was unlawful coercion, so there’s this famous case from the Seventh Circuit backpage.com v. Dart, which involves the Cook County Sheriff putting pressure on credit card companies to pull out a providing services on backpage.com because he thinks that it’s involved in illegal sex trafficking. The court finds that this does violate the First Amendment, but on the stand, a representative, I think it was from Visa or MasterCard, one of the credit card companies said, “Oh, no, no, we weren’t threatened. We weren’t threatened.” And the judge, Judge Posner says, “What are they going to say? It’d be so bad if they say our business decisions were dictated by those local sheriffs. The platform simply saying we would’ve done it anyway, doesn’t give me a lot of comfort.”

That said, I think the general problem with the opinion is something I alluded to earlier, which is that the court’s not at all considering the downsides of having an incredibly broad rule against jawboning. It’s not trying to balance the interests that are harmed by preventing government officials from talking to platforms or encouraging platforms. The result is, or in a way what you can see this in that it’s just saying significant encouragement is any encouragement, especially if it’s repeated. When I’m thinking about jawboning cases, sometimes I think it’s useful to make a comparison to the entrapment cases in the criminal context. When a police officer puts too much pressure on a private person to violate the law, they go undercover and they pretend that they’re a criminal too, and they say, “Hey, let’s go rob this store.” You can have a defensive entrapment if you can show that the coercion is too much.

And as you can see, the kinds of considerations here are similar in that there’s certain forms of government coercion and pressure on private parties that we might think are really bad because it’s not fair to put someone in jail for something that they were pressured into doing by an undercover cop. On the other hand, we want there to be the ability to figure out when people are violating the laws, and sometimes the best way to do that is to have undercover activity and undercover agents collaborating with private people to engage in law-breaking, so there are interests on both sides.

In the all which long wind up to say in the entrapment context, courts differ about the test that they’re going to use to figure out when entrapment goes to far. There’s an objective test which just looks at what the government did, and it doesn’t care about whether or not the person who is entrapped would’ve done it anyway, because they think that there’s certain forms of pressure that the government shouldn’t engage in full stop. Then there are other courts that no, they look at, it’s a subjective test. You can only find entrapment if you can prove that the private person wouldn’t have done it anyway. I’ll just say in the entrapment context, there is this big debate about how much that matters.

What’s so different about those cases from this case is in both lines of precedent, the test for entrapment is narrow. For good or for ill, but there is lots of government pressure, lots of forms of government complicity with private people, undercover police officers working with people who they think are going to engage in crime that doesn’t constitute entrapment because the courts are trying to balance interests, and here there is just none of that at all. You see it in the breadth of the injunction, which essentially stops government officials from across a huge swath of the executive branch from talking to private people in the platforms about any harmful speech on their platform at all, and doesn’t do any balancing of interests.

Evelyn Douek:

I want to come back to the injunction and the interests that are at stake and how we should think about those, but before we do and leave this very wonky and doctrinal corner, you mentioned that the Bloom test has generally been considered or interpreted very narrowly. Including in the social media context specifically and that one of the incredible things about this opinion is its departure from that, and so I think it’d be useful to place this judgment in contrast to some of those other cases that have been about social media platforms and governments relationships and how other courts and other circuits are thinking about this because of course, ultimately we’re going to come to this way down the track, but one of the things that we’re seeing is the divergence of the law in different circuits, and that is setting up a nice potential Supreme Court petition at some point. How are other courts thinking about this?

Genevieve Lakier:

Well, the biggest contrast or the easiest way to see the contrast is to think about this decision versus the decisions that are coming out of the Ninth Circuit, so for obvious reasons, many, many of the jawboning cases that we’ve seen have come out of districts, and then the Ninth Circuit Courts of Appeal. District Courts in the Ninth Circuit from California, and then the Ninth Circuit Court of Appeal, which covers the West Coast. The Ninth Circuit test seems to be, there was a few cases handed down I think, last year, but one of the most recent is the Kennedy v. Warren case, which involved a claim by Robert Kennedy Jr. Who’s talked about in this case too, that Elizabeth Warren sent, Elizabeth Warren violated the Constitution when she urged the platforms to take down some of his vaccine skeptic speech.

And the Ninth Circuit held that there’s no jawboning because there was no evidence. Well, first Senator Warren is a legislator and doesn’t have direct regulatory authority over the platforms, and so there’s this strict test, which would also be a problem here of who has regulatory authority. Second, this language that it takes from, I think actually from Bloom, which says that unless you direct or command the result, there’s no jawboning, and there’s just no evidence of a close enough connection between the government actor, in that case, Warren, actually decide saying, “In this case, you shall do this.” And what happened? And so just incredibly narrow test that was extremely easy for the Ninth Circuit to dismiss.

The facts are different here. It’s true. What makes this complicated is that there are a few emails in which it does look, maybe, I don’t know, people in the White House are trying to direct a particular result with respect to particular Speech Act, but then there’s tons and tons of other cases in which as you said, Evelyn, there’s nothing like that we’re not even in the same universe as that kind of behavior. The court, while purporting to apply this Bloom test, doesn’t distinguish between them at all, which is again just incredible.

Evelyn Douek:

And just to even make the contrast even sharper, there was also a Ninth Circuit decision earlier this year, Han Levy Weber, where it was about the Twitter partnership portal that it has in the context generally where it allows third-parties to flag content to them for review. Like a trusted flagger program where certain people’s flags get higher priority review, and it gives access to this portal to a number of different third-party actors, and some of whom are government officials, some are other civil society actors or whatever, but people who might have specific or particular knowledge or expertise that might make their flags generally more trustworthy or more valuable to the platform, because anyone that works in the space knows that flags can generally be very noisy signals of where the content is actually violating.

And so this was in the context of an election, the Secretary of State of California was staff as we’re using this portal to flag election misinformation to the platform and the court dismisses this and says that this is clearly not a violation of the First Amendment. This is not the kind of entanglement or coercion or it’s joint participation or meeting of the minds that the First Amendment generally thinks about in this context. That exact action is also in this district court opinion. Those exact facts of government officials using partner portals to flag things to platforms. It’s not all these emails, it’s also these more formalized mechanisms that platforms have been setting up, which do raise all sorts of problems that in terms of transparency and what’s going on, but also as we’ve talked about, they aren’t always successful. Not all the flags are always accepted, and the Ninth Circuit finds that very unconcerning, whereas reading this district court opinion, what exactly was the wording? That the biggest assault on free speech in this nation’s history. And so just a completely different perspective on the exact same behaviors.

Genevieve Lakier:

I might be a bit like Goldilocks, never satisfied with the temperature of the porridge here though, because I also, I really don’t like that opinion though. Handy opinion. I think that goes too far in the other direction, because wasn’t I… I don’t remember the specifics. It’s been a while since I read it, but in that case, it was a very high acceptance rate for the flags. It was something like 90% were taken down after they were being flagged. I don’t know, I think a test, like the test that the Ninth Circuit relied on there and has relied on in the other cases where you have to show that the government actor actually directed or controlled the result, I think it allows too much evasion.

So long as the government phrases their actions as a request and there are some instances in which the platforms rejected. It seems like the test is not satisfied, no state action, no First Amendment problem. We might imagine that in practice that could result in a lot of government interference in the decision making of the platforms. Again, I don’t know. I don’t know what the motivation is. There’s so much we don’t know, but that felt to me in that case you didn’t even get to discovery because the court said, “As a matter of law, you haven’t alleged enough to show a First Amendment violation, so we’re not going to find any more facts.” That’s too narrow but it’s true that this district court opinion is too broad.

Evelyn Douek:

I don’t remember the exact acceptance rate in that case either, but I have conflicting feelings about the acceptance rates, because on the one hand, that could actually just mean government actors are being very cognizant of their First Amendment limitations and are only really genuinely flagging things that are within the platform’s rules and are explicitly here is the wrong time and place of voting or whatever it happens to be that the rules are about, and they’re trying not to overreach, and so they’re not abusing that power. There’s multiple ways to think about it and ultimately, and again, this is going to be in that second podcast that we record sometimes in terms of thinking about better ways of thinking about this problem.

But to my mind, a better way of thinking about it is like, “Well, what procedures and mechanisms does the platform have in place when it’s reviewing those flags to show that they are reviewing it in a way that the government portion of it is not particularly relevant to their assessment?” They are reviewing that content. Maybe it’s flagged in one portal, but it’s given to a reviewer who doesn’t necessarily know that the government actor was the person that flagged it and they’re conducting an independent review. Then we have transparency around what has been taken down as a result of government flagging and things like that. Those are the kinds of procedural mechanisms I’d be much more interested in seeing rather than we get into a place where we’re looking at sheer numbers, but I completely agree with you that a lot of these cases.

Part of the problem has been a lot of these cases have been generally pretty unconcerned about these relationships, this new problem that we have of these relationships between platforms and governments. On the one hand, you have this district court saying, “This is the greatest assault on free speech ever.” And these other courts not really blinking an eye and I similarly think that that’s a problem. Why might they be doing that? What is the impetus behind why are those courts so unconcerned? We normally think of courts as really taking a broad view of the First Amendment and the First Amendment being weaponized to intrude into a whole bunch of areas. Why not here?

Genevieve Lakier:

I think there is a wide widely shared consensus among a lot of judges, and I also think First Amendment people, platform people generally that in information sharing, this kind of communication between government officials, particularly government experts of certain kinds, so people in the CDC, the kinds of individuals that are being defendants in this lawsuit specifically. CDC people are experts on elections, public health people more generally, that information sharing between them and the platforms is really important because it is true that there’s so much harmful and misleading speech circulating on the platforms. I also think there’s this view that the old forms of gatekeeping and ensuring that only credible information or reasonably credible information made it into broad public dissemination, which was relying on the traditional legacy news media and professional ethics and response, all the old forms of gatekeeping, those are less important, and so we have a much more democratized public sphere.

But given a much more democratized public sphere, there’s just a lot more crazy, wrong, harmful speech out there. You might think that civil society has something to say about what kind of speech is dangerous and doing bad things, but you also might think that the government officials, the government bureaucracy, has a lot of expert knowledge and information about what’s harmful speech, what’s misleading speech. Some of the allegations in this case I think make that very clear, illuminate that. In some of these cases, the platforms would go into the CDC because in quickly changing information environment. COVID had just hit, people were trying to figure out what was going on, information, recommendations about what to do, were changing. The platforms didn’t know how to distinguish between good and harmful speech about COVID and what was the scientific consensus on remediation techniques and masks, and so they would go to the CDC to ask for information.

While we now know the CDC made mistakes, wasn’t perfect, you might think that still the CDC has much more information than Mark Zuckerberg or anyone at Twitter. However hardworking they are, that’s just not their area of expertise, and so to cut off information sharing between the platforms and these kinds of government expert agencies is harmful to the overall health and accuracy of the public sphere. And so I think the courts are just really worried about the costs of having too broad a First Amendment rule.

I think also there is this worry about government speech interests too. The First Amendment protects private people, it doesn’t protect government actors, so government speakers, government officials don’t exactly have a First Amendment right. But courts, particularly in the last few decades have recognize that it’s important for government officials, government representatives to be able to speak about policy matters and express their viewpoints. That’s how democracy works. You can’t have a Democratic government if you don’t have a government that is able to take policy positions, and so we really want to protect the ability of government actors to speak.

And in this context too, I think courts recognize that government criticism of the platforms is part of the process of democratic oversight of these really powerful private corporations. It would be harmful to democratic interests for government officials never to be able to say the platforms are doing a bad job in this way or that way. That is a form of bringing to bear some kind of public pressure on platforms when they’re acting in a way that seems contrary to the public interest. Now maybe in some context that can go too far, but some amount of that speech seems really important democratically into having a healthy public sphere.

And so I think it’s because of their concern with these very strong and serious interests. Interests in expert knowledge, interest in democracy, interest in government expressive freedom that the plat… And I think also just this real worry about state action, about not expanding the First Amendment too far, not having courts second guessing every time a government official talks to a platform. I think the courts just do not want to get involved in this that has led them to have this incredibly narrow rule, but the result is that it has meant, even in cases where we might think that there’s something really concerning going on, it doesn’t seem like the courts are worried.

Evelyn Douek:

It is remarkable by contrast how little concern this district court opinion has for the importance of government speech. When it’s talking about Fauci doing live streams on Facebook or statements from the press briefing room, or even Biden saying Facebook is killing people. I think about there are constantly government actors criticizing the New York Times or Fox News for that matter, criticism of the way that speech intermediaries operate is stock standard in politics. It’s something that happens, and so lumping those kinds of things in with these other forms of hard contact was particularly, I think, surprising given the difference.

Genevieve Lakier:

I think the district court’s view, and this is, maybe partly a product of a really polarized information environment. There’s a line I think in the opinion where it says it is telling or it’s important that all of the speech targeted here is conservative speech. Now, of course, all the speech being targeted is conservative speech because the plaintiffs chose to only identify instances in which conservative speech is being targeted. That seems like a strange way to think about the reality, but I think it reflects the district court’s view that we don’t need to worry about the harm to expert knowledge because what’s at stake here is just viewpoint discrimination. It’s just a targeted campaign of suppression.

If this were really about making sure the truth is out there, it would fall in a more equitable way on conservative and liberal views, et cetera. I think that’s what the district court is saying, that I don’t have to worry about balancing these interests. I don’t have to care about the interests on the other side because the government’s doing something so bad, but of course, it only thinks that because it has a very simplistic view of what’s at stake here.

Evelyn Douek:

Okay, so let’s talk about the academics. Significant parts of the opinion are dedicated to finding that the government was so entangled with the work of certain academics, including many of my colleagues here at Stanford, and one person particularly well known to subscribers of this feed, Alex Stamos in his capacity as director of the Stanford Internet Observatory, so as to make their work a violation of the First Amendment. Just a quick list disclaimer here, I am at Stanford, but I was not involved in any of the work of the Stanford Internet Observatory, and so I’m not, wasn’t involved in any of the work that’s mentioned in the opinion. I’m speaking purely in my academic capacity and not affiliated with this work, but I’m clearly also not impartial about this part of the opinion or this part of the movement in this area more generally against academics and academic work. Genevieve, what does the opinion say about academics and academic work here?

Genevieve Lakier:

I have to say this is another part of the opinion that is jaw-dropping or incredible, because first of all, none of the facts in the opinion, none of the facts that the district court invokes suggests that anything improper happened at all on the part of the academic researchers. Although the court manipulates the facts I think a little bit or is carefully and selected the facts and suggesting that there’s some kind of conspiratorial agreement between the government and the private organizations like the SIO to collaborate to evade the First Amendment.

And here again, I think this is, maybe this is the best example of the court failing to think about the countervailing interests, so it says, because we know that there was some collaboration be between these private researchers and the government insofar as there was information sharing and the researchers communicated with the government and then communicated with the platforms about election misinformation or other harmful speech that establishes that there is some kind of conspiracy and a violation of the First Amendment rights and jawboning. By the way, there’s, as far as I can tell on the part of the government officials, no evidence of not even coercion, but significant encouragement. The Stanford Internet Observatory is doing its own thing, but there is connections to the government.

In finding and sweeping this all into the same bundle of bad stuff that includes the emails we talked about earlier and the fake account with Dr. Fauci and not distinguishing between the two, the court seems to have no concern for the free speech interests and the important research interests that are might be chilled by its conclusion that this is a violation of the First Amendment. One might think as a First Amendment person that it’s a much better situation from a First Amendment perspective, from a democratic perspective to have experts, civil society agencies that are thinking about harmful speech on the platforms and then speaking to the platforms about that speech that is not government controlled or government staff but is private, then having all this stuff have to be delegated to the government.

Although as I said earlier, we might think that the government has expert information and we want there to be sharing between the government and the platforms about that information, it’s true that we don’t want the government to be too involved in the business of these private speech intermediaries. The whole reason we have a privately owned and operated public sphere, privately owned public sphere is because we think it’s important for there to be some kind of separation between the government and the private public sphere, privately owned public sphere.

These organizations that are trying to do the work of promoting the public interest by flagging bad content for the platforms seem to be supporting the aim of a separation between the government and the private privately owned public sphere, but the court doesn’t recognize that at all and then in the end, because it’s smushing all of these different, very, very different phenomena together issues in injunction that appears likely to have a significant chilling effect on these private researchers who are not government actors, and the opinion doesn’t say they’re government actors and they’re not defendants in the suit, so there’s no allegation that they have been turned into government actors and yet still now government officials are enjoined from being able to talk to them about any harmful speech on the platforms as well.

I think it’s maybe the best example of the district court’s utter failure to think about the very important free speech interests that are threatened by the way in which it’s going about the First amendment analysis here.

Evelyn Douek:

This is one of those moments where this opinion and this injunction notionally about protecting free speech is being used to cast doubt on the academic work of independent researchers who are exercising their own first amendment rights in the work that they do, including talking to the government but including talking to platforms in a way that is as having chilling effects for sure and seems a bit topsy-turvy land when we think about how this is supposed to work.

Let’s talk about the injunction. We’ve noted that it’s extremely broad and we can talk about what exactly it prohibits, but also I think interestingly what it doesn’t prohibit and what it exempts from its prohibitions, which are in some ways also extremely telling because they recognize that there are certain times when we might want government actors to be talking to platforms about certain things and there are times where this kind of contact is legitimate and in the public interest, so in some ways the injunction itself highlights the very difficult questions here in the different equities, but of course it doesn’t really grapple with them or deal with them in any meaningful way. Tell us about in the injunction.

Genevieve Lakier:

The injunction follows I think pretty straightforwardly from the opinion. The opinion says essentially any communication to or about harmful speech on the platforms made by government officials or government officials working in some loose collaboration with these private researchers violates the first amendment. And so the injunction says in order to prevent continuing violations of the first Amendment, you may not encourage, induce, flag, and I think the language is in any manner speech that is protected, speech that is harmful or that you want to be suppressed or reduced. I think basically means is incredibly broadly worded, basically means that all of the agencies identified in the injunction and all of the employees of those agencies from now until it ends or is changed really cannot talk about harmful speech on the platforms in any way.

And it’s not obvious to me that it is just referring to direct communication between the government actors and the platforms. It seems like it’s also talking about public speech, so speech that could have the effect indirectly of encouraging the platforms to take on the speech seems like it’s covered by the injunction as well. As you said, then there are these exceptions including my favorite, which is the exception for that you can alert the platforms to threats to national safety and security when it seems to me like all of the government defendants involved in the suit, wouldn’t they say that’s the only thing we’ve been doing. We’ve been identifying threats to national safety and security, threats to election safety, threats to public health, which is a form of safety and security. What else have they been doing?

And so there’s a way in which the injunction seems like it stops all communication between the government agencies and the platforms about bad speech, and then there’s a way in which maybe it doesn’t change anything at all, depends on how you define the exceptions. The injunction doesn’t provide any definitions whatsoever, so I really do not understand how the injunction is supposed to work. Who figures out the differences? I think it’s probably intended just to stop all speech because no one knows what it means, but it’s just, again, as you said, failing to… Acknowledging in some way, that’s a win. Acknowledging that there are countervailing interests but not grappling with how to reconcile them at all.

Evelyn Douek:

Right. The way the injunction is supposed to work is it’s supposed to, especially in light of the opinion that preceded it, just chill broad kinds of communication of the type that we’ve talked about in the opinion between platforms and the government actors, but also to have this plausible deniability of, “Oh no, but we didn’t mean the times when it’s really, really necessary.” But that chilling effect like the government actors going to are on the side of caution, and we’ve already seen that the Washington Post had reporting yesterday about how certain departments are canceling meetings that they had with the platforms. It’s having its intended effect even if the exceptions are notionally there.

I think one of the things that’s just really notable here is I think that there are genuine problems that social media platforms raise for jawboning and in this area that the doctrine doesn’t really deal with. We have talked a lot about how a lot of the precedent is around this coercion and threat standards. Bantam Books is about when the government actor threatens a private actor with bad consequences to remove speech. That doesn’t necessarily fit well with the situation that we currently have where these platforms do have these relationship with the government, but they’re not always, sometimes they are platforms set up these channels and want government knowledge and information to be shared with them.

And there’s this regularity and formalization of these relationships that I think is somewhat unprecedented or at least hasn’t been dealt with in the cases so much. There’s also the scale of content moderation and the problems that it raises. I think it makes government expertise particularly important. One of the things that I’ve been talking about when I’m talking to reporters about this is in terms of local election misinformation, platforms aren’t going to be great at policing here’s some posts tweeting the wrong time or place of a polling booth. That is the kind of thing where local officials are going to be particularly well suited to flag that to the platforms, and so those kinds of relationships that can be beneficial and we can understand why that’s being set up.

But it’s that consensual relationship or that isn’t necessarily always based around threats, although we have also talked about times where there are threats involved, that isn’t dealt with well in the doctrine, but I do think for all of the reasons that we’ve talked about. That doesn’t mean, “Oh look, we’ve never talked about this before in the cases and so the First Amendment shouldn’t think about it.” I think the First Amendment needs to think about this in meaningful ways, but it hasn’t been yet and I’m just curious, do you agree with that synopsis of where we stand? Where have this genuine problem that the First Amendment’s not dealing with and in some ways this district court is trying or it is acknowledging the problem but not really dealing with it in a meaningful way?

Genevieve Lakier:

I agree with you that the First Amendment cases were developed with a very different model in mind for government, private speech platform relationships in mind that things have changed on the ground and that the doctrine is failing to grapple with it and really should. I think the general relationship between government regulators and private speech platforms up until the social media era took one of two forms. Either you have local sheriffs, local police officers going off to local books, sellers or local billboard owners to take down particular kinds of speech that they find really egregious. This is episodic, this is very targeted, it’s very viewpoint discriminatory. It’s one-sided. The booksellers and the billboard owners they did not invite any kind of attention from the government, but they’re doing something that’s violating the local norms and the government’s going after it, and the question the courts have to deal with is does this go too far or is this within the power of the local government?

And then you have these, I think largely chummy relationships between the large legacy media institutions and federal regulatory agencies and executives. There are cases that describe FCC or people from the White House picking up the phone and talking to the CEOs of NBC or ABC or the large newspapers and trying to get them to do particular things or not do particular things, but this also is, I think relatively infrequent and it’s based on everyone being members of the elite, a lot of personal relationships, a lot of shared agreement about what we’re trying to achieve and how does the First Amendment deal with those kinds of, I think what’s true in many industries, these buddy-buddy close, pervasive social and political ties between the industrial elite, the owners of the large businesses and the government. That’s the the old way in which there was government oversight or influence coercion of these private speech platforms.

And now we have this just very different situation. From a First Amendment perspective in some ways much better and also much worse, just very different in that, and you can see this in the complaint, this very quite bureaucratized, expansive, regular form of information sharing, meeting, coordination, conversation between these platform officials. We’re not talking about just a few really well-placed elite actors. We’re just talking about employees of the platforms in lots of different capacities and then representatives from a huge array of different agencies and the White House having, it was amazing. The opinion’s really interesting for what it reveals about how regular and how frequent this kind of information sharing both between members of the White House and then other federal agencies and I would imagine also state officials are.

And so what we have here is a bureaucratized system that looks, and I think it fits into this broader story in which the platforms have taken on a quasi-governmental role or have been asked by the government, pushed by the government to take on a quasi-governmental role when it comes to things about, and not just speech, when it comes to security and privacy, enforcement of copyright rules in so many different ways the platforms are doing things that look quite governmental and they’re doing it in close collaboration with the government and that required a lot of information sharing. In some ways that seems terrible from a First Amendment perspective, if we have this view that there should just be a strict separation between the government and the private sphere because it allows too much manipulation, and I think we should be worried about that. That’s definitely concern.

On the other hand, there is something comforting to me about a much more bureaucratized formalized system that’s not just about elite connections, because so long as there’s adequate accountability and oversight as you were saying, of the mechanisms by which we do in this, some kind of safeguards to ensure that it doesn’t go awry or that there’s not too much government pressure. It seems less susceptible to one off to particular speech because you don’t like it. Now of course that still seems to happen. I think some of the emails that are talked about in the opinion suggest that there’s still that capacity to push on the platforms to remove particular speech that a government official doesn’t like. Maybe there’s more going on there and it’s more understandable, but I’m not trying to be a Pollyanna about the current system.

But I do think that a more bureaucratized, more pervasive form of platform government relation has some upsides because it is more subject to, we can think about rules to govern it in some kind of accountable way, democratically accountable way, but also maybe pushes us, pushes the doctrine or requires the doctrine to think in new ways about the harms here because when it’s just a local police officer intimidating a bookseller, the harm is the way in which you think about what’s the bad thing happening here is that it’s threatening, it’s a threat. It’s all about threats.

Today I think we might think that there’s a lot of government involvement and maybe even government manipulation of the information environment that doesn’t take the form of threats. It takes the form of collaboration or information sharing, coordination of some kind. The cases just haven’t really been focused a lot. They tend to, I think we’re seeing more efforts to articulate this as a First Amendment problem, but the conventional way of thinking about when jawboning violates the first amendment is when it’s threatening, when it’s coercive. Even in this opinion, the court goes back and forth between the language of significant encouragement and the language of encouragement, and I think that there’s a distinction. I think we might think, I’m not sure, but we might think that there are First Amendment problems even when there isn’t coercion, even when there isn’t threat, but there’s too much collaboration or too much involvement, too much manipulation by the government through these forms of information sharing.

But the doctrine absolutely hasn’t figured out how to deal with this new kind of pervasive, routinized, somewhat bureaucratized form of platform government cooperation, but I agree with you a hundred percent. It’s very new.

Evelyn Douek:

Of course, this is a problem for First Amendment doctrine, but it would be remiss of me to be too parochial before we close completely and not note that this is a problem globally and this is a global problem. That in some ways you can test your intuitions for how you think about this for thinking about, well, how do we feel about the same rules applied in different countries? In fact, one of the last long form conversations we had on this podcast feed was in the context of Taamneh, Twitter v. Taamneh, and there was a conversation in that case about, because that case obviously was about whether Twitter could be held liable for aiding and abetting terrorism. There was a context about what level of sufficient knowledge do they have to have before that can be found? And there was this exchange that happened in the oral argument where there was well, if the Istanbul Police came to Twitter and said, “Hey, these accounts are really problematic accounts, you need to take them down.” And Twitter refused to take them down, that would would make them culpable.

But of course that is exactly that highlights the problem, because it is both yes, we can understand why we might think Istanbul Police might know particular accounts, but that exchange was happening also in a time where the Turkish government was literally locking up journalists for spreading fake news and misinformation, and so it highlights both the promises and perils of those kinds of relationships and the ways in which jawboning can be problematic. We’ve also seen it in the context of the oversight board had a case about the UK Metropolitan Police flagging to Facebook instances of drill music, which I haven’t listened to myself, but I understand is a particular kind of music that’s very popular in the UK and also popular under certain racial, to in particular with certain racial communities, and so the UK police flagging this to platforms had all of these racial dimensions and discriminate problems as well, and the oversight board explicitly talked about the importance of transparency and the problems of these relationships as well.

And so I think those are just two examples to highlight how we should think about this both more broadly and understand that this is a global problem because we’ve been talking a lot about the First Amendment, but of course these platforms are global and they need to deal with this with lots of different governments.

Genevieve Lakier:

Agreed, a hundred percent.

Evelyn Douek:

Okay, so what do we think will happen here? What’s next? What do you think? We’ve already seen the government actors have filed a notice to appeal and they’ve filed a request to stay the injunction pending that appeal. This was a district court judgment. Play this out for us very briefly. I’m not asking you to stare into a crystal ball, but what should listeners be watching for expecting?

Genevieve Lakier:

My prediction, but again, I’m not a talented court watcher, but let’s just say that because this is the Fifth Circuit and the Fifth Circuit has shown a willingness, it’s shown maybe even sometimes an eagerness to deviate from the status quo doctrinally when it comes to the first amendment. We saw that in the NetChoice cases where it upheld Texas’s new social media law.

Evelyn Douek:

Which of course are not a form of illegitimate government influence on what platforms might or might not carry on their services.

Genevieve Lakier:

Let’s leave that to another podcast.

Evelyn Douek:

Right.

Genevieve Lakier:

That I think it’s quite likely that the Fifth Circuit is going to affirm this in some form or another, even if it is, as I say, as I said earlier, pretty significant departure from precedent, but it seems likely that they would. The most likely outcome I think is that they’re going to affirm the injunction but narrow it in some way just because it is, as we suggested, really poorly drafted, very unclear, very expansive, so I think it’s likely that the First Amendment is going to affirm, but no. If that is what it does, then we have a circuit split. If we have a circuit split and a pretty significant circuit split about a pretty significant issue, because if the Fifth Circuit affirms this in any way, it’s going to mean that it’s adopting a test that looks incredibly different from the test that the Ninth Circuit or has relied on, and so we have this very significant circuit split.

Seems then that surely the Supreme Court is going to have to weigh in some form or another, and that could be really interesting as I think we both suggested the law, the constitutional law of jawboning needs some updating. Whether the Supreme Court is the institution to do a good job of updating. As Justice Kagan said, there are no experts on the internet, and this is all very complicated and difficult as we’ve been suggesting. I don’t know what to expect from that, but that’s what I’m watching for.

Evelyn Douek:

Good things. Of course. We expect good things from the Supreme Court. Do you know what I mean?

Genevieve Lakier:

Always. Always.

Evelyn Douek:

All right, so we’ve gone very long here, longer than intended. I think that’s probably a testament to how intricate and interesting a lot of these issues are. It’s also perhaps an unintentional filter on people. Those who are looking for extremely hot and spicy takes do not come to an hour and a half long podcast between two law professors, but we will leave it there. This has been your moderated content episode. This show is available in all the usual places, including Apple Podcasts and Spotify. Show notes are available at law.stanford.edu/moderated content, and that’s where we’ll be posting a transcript as well. Thank you to the wonderful producer Brian Pelletier for putting this together, and special thanks also to Justin Fu and Rob Huffman. See you next week.