Free Speech on Facebook and Twitter
by Michael C. Dorf
Two social media platforms have adopted very different approaches to political advertisements. Facebook has a general policy of fact-checking and removing "false news and other types of viral misinformation, like memes or manipulated photos and videos," but exempts politicians, including politicians who appear in ads on Facebook. Twitter, formulating its policy in deliberate contrast with the Facebook policy, will stop running political ads entirely.
Both Facebook and Twitter justify their respective policies by reference to norms of democracy and free speech. Let's consider the merits of these justifications.
Neither Facebook nor Twitter is a state actor, so neither is governed by the First Amendment or other constitutional norms (aside from the 13th Amendment's prohibition on slavery). That's not only a fundamental principle of constitutional law. It's also essential to the multinational business of social media platforms. By law, Facebook and Twitter must censor hate-speech in many of the countries in which they operate; the government must not censor hate-speech in the US; so if Facebook and Twitter were state actors for US law purposes, they would face incompatible legal demands--or they would have to create separate country-specific social media ecosystems.
Although the Constitution does not regulate Facebook, Twitter, or other platforms, federal statutes do. One statute, the Digital Millennium Copyright Act, imposes obligations on Internet platforms to take down copyright-infringing material upon notification. But platforms have no general obligation to take down content that makes false assertions of fact. Even defamatory false speech may remain up on an Internet or social media platform, because Section 230 of the Communications Decency Act of 1996 says that such platforms are not treated as the speaker or publisher for (non-copyright) liability purposes. The Act pre-empts state laws to the contrary.
Accordingly, Facebook, Twitter, and other platforms could adopt a policy of laissez-faire with respect to fake news. The fact that they have not done so reflects some combination of idealism and business judgment in response to public pressure. Yet while both Facebook and Twitter justify their respective (quite different) approaches as aiming at promoting democratic values, each approach would be unconstitutional if undertaken by a government actor.
In recent testimony before Congress, under questioning from Congresswoman Alexandria Ocasio-Cortez, Facebook's Mark Zuckerberg was less than pellucid in his explanation of the scope of the company policy of exempting politicians in political ads from its fact-checking protocol, but his justification for the policy (whatever its scope) was straightforward enough: Zuckerberg said people should be allowed to "see for themselves" what politicians say and draw their own conclusions, rather than have the company decide what they can see.
That rationale sounds broadly similar to the rationale that Justice Kennedy gave for protecting political lies in United States v. Alvarez. Issuing a warning of an Orwellian future, he wrote for the Court: "Our constitutional tradition stands against the idea that we need Oceania’s Ministry of Truth." So too, one might say, we do not want giant companies like Facebook acting as powerful censors, even if the penalty for false speech is only its removal, rather than a rat to the face in room 101.
Yet note that Facebook's policy, if adopted by a state actor, would violate the First Amendment, because it is speaker-based. A long line of SCOTUS cases treats distinctions among speakers as pernicious in the same way that distinctions based on the content or viewpoint of speech. One might think that the Supreme Court misapplied that line of cases when, in Citizens United v. FEC, it condemned a distinction between campaign speech by natural persons and such speech by corporations, but still agree with the general proposition that speaker-based distinctions often do amount to censorship. As the Court put the point in Citizens United: "Quite apart from the purpose or effect of regulating content, ... the Government may commit a constitutional wrong when by law it identifies certain preferred speakers."
That constitutional wrong would very much be committed by a government rule that privileged speech by politicians over speech by ordinary citizens. True, such a rule would not stifle all dissent: People running for office would be permitted to criticize incumbent politicians. However, a rule favoring speech by politicians will naturally tend to squelch dissent and undercut democracy, in which citizens are entitled to tell politicians what they think rather than simply listen quietly.
To be sure, the Facebook policy does not forbid speech by ordinary citizens; it merely disfavors false speech by non-politicians. Yet that distinction seems inadequate. Underwriting Alvarez is the recognition that government ought not regulate the truth or falsity of political speech, regardless of who the speaker is. The case struck down the Stolen Valor Act in general, not just as applied to politicians.
Accordingly, there can be little doubt that Facebook's policy would be unconstitutional if adopted by the government. Does it follow that the policy contravenes basic free speech principles of the sort that Zuckerberg wants to respect voluntarily?
Not necessarily. Some free speech norms for the government do not translate to private actors. For example, a government cannot decide which opinions to stifle and which to amplify, but a newspaper publisher committed to principles of free speech can and will make such judgments in deciding whom to hire as an Opinion writer, what op-eds to publish, and what official editorials to run. Even a newspaper publisher who wants to provide a wide range of opinions will typically deem some opinions on some subjects beyond the pale. The NY Times took a great deal of flak when it hired climate skeptic Bret Stephens to provide "balance." No one at the Times would have thought to hire a Holocaust skeptic, a moon landing skeptic, or a Pizzagate believer to provide balance on those issues, even though the government may not censor Holocaust denial, moon landing skepticism, or Pizzagate conspiracy theories.
Thus, it is at least possible that Facebook could justify its policy on grounds that free speech norms applicable to the government do not apply to it in the same way, even assuming that Facebook wishes to comply with free speech norms.
That's possible, but in practice the claim fails, because Facebook's policy makes little sense. AOC's cross-examination of Zuckerberg showed that he hadn't thought it through very well. And in practice, the policy has already been applied in a way that shows how it favors mainstream views. Taking inspiration from one of AOC's questions for Zuckerberg, Adriel Hampton created a fake ad appearing to show Sen. Lindsey Graham saying "I support the Green New Deal." That didn't fall within the exemption for politicians, because it was speech by Hampton, not Graham, so Hampton registered as a candidate for Governor of California. Facebook still censored Hampton on the ground that he wasn't really a candidate. Maybe that's right, but is Facebook really well positioned to say who's a genuine candidate for office? In 2016 millions of people thought Donald Trump was running for president only as a publicity stunt.
Accordingly, the Facebook policy is misguided and contrary to principles of free speech. What about the Twitter policy? It too would be obviously unconstitutional if adopted by a state actor. The government cannot ban political advertisements. Might there be a justification for Twitter's policy coming from a private actor?
Sure. One justification would be dissociation from controversial speech. Suppose the private owners of a sports arena sell advertising space (as nearly all of them do). They might be perfectly happy running ads for cars, beer, office supplies, and numerous other proprietors of goods and services but unwilling to sell space for political ads. Michael Jordan apparently never said, "Republicans buy sneakers too," but if he had, that would have been a legitimate point about business. The public might not distinguish between the message and the platform, so the owners of the platform have a legitimate interest in not selling political ads--a dissociation interest that is unrelated to censorship.
Consider Games 5 and 7 of the recently completed World Series. When Donald Trump attended Game 5 in Nationals Park, the home fans booed and chanted "lock him up" when the Jumbotron panned to him. When the fans holding a Game 7 Nationals Park watch party of the tv broadcast from Houston saw a Trump ad, they booed again. Good for the Nats fans, I want to say, and so I just did. However, one could see how the Nats owners wouldn't want any fans to feel uncomfortable based on politics at a game, so from their perspective, Trump's attendance and the showing of the ad were unfortunate--although Nationals Park is publicly owned, so the First Amendment applies of its own force, so the owners were somewhat more constrained in their ability to prevent political advertising than they would have been if the Nats played in a privately owned stadium.
In any event, Twitter does not invoke the dissociation interest. Instead, Twitter CEO Jack Dorsey has said that he bans political ads but not Tweets because politicians should earn views on Twitter, not buy them. Is that a good rationale?
Maybe. Dorsey might think that money from the actual marketplace corrupts the Holmesian marketplace of ideas. If so, his view has much in common with the principles behind campaign finance regulations that limit both contributions and spending. To be sure, the Supreme Court has invalidated this sort of rationale in its First Amendment case law, but to repeat, that case law does not apply to Twitter.
Ultimately, whether one thinks the Twitter approach is justified depends on what one thinks of the Court's campaign finance cases. Twitter's approach does not rest on the notion that it is situated very differently from the government with respect to a broad conception of what counts as the corrupting influence of money on politics. Twitter's approach instead rests on the notion that the Supreme Court has misunderstood the First Amendment. Many readers may well agree with Twitter, not the SCOTUS. I know I do.
Two social media platforms have adopted very different approaches to political advertisements. Facebook has a general policy of fact-checking and removing "false news and other types of viral misinformation, like memes or manipulated photos and videos," but exempts politicians, including politicians who appear in ads on Facebook. Twitter, formulating its policy in deliberate contrast with the Facebook policy, will stop running political ads entirely.
Both Facebook and Twitter justify their respective policies by reference to norms of democracy and free speech. Let's consider the merits of these justifications.
Neither Facebook nor Twitter is a state actor, so neither is governed by the First Amendment or other constitutional norms (aside from the 13th Amendment's prohibition on slavery). That's not only a fundamental principle of constitutional law. It's also essential to the multinational business of social media platforms. By law, Facebook and Twitter must censor hate-speech in many of the countries in which they operate; the government must not censor hate-speech in the US; so if Facebook and Twitter were state actors for US law purposes, they would face incompatible legal demands--or they would have to create separate country-specific social media ecosystems.
Although the Constitution does not regulate Facebook, Twitter, or other platforms, federal statutes do. One statute, the Digital Millennium Copyright Act, imposes obligations on Internet platforms to take down copyright-infringing material upon notification. But platforms have no general obligation to take down content that makes false assertions of fact. Even defamatory false speech may remain up on an Internet or social media platform, because Section 230 of the Communications Decency Act of 1996 says that such platforms are not treated as the speaker or publisher for (non-copyright) liability purposes. The Act pre-empts state laws to the contrary.
Accordingly, Facebook, Twitter, and other platforms could adopt a policy of laissez-faire with respect to fake news. The fact that they have not done so reflects some combination of idealism and business judgment in response to public pressure. Yet while both Facebook and Twitter justify their respective (quite different) approaches as aiming at promoting democratic values, each approach would be unconstitutional if undertaken by a government actor.
In recent testimony before Congress, under questioning from Congresswoman Alexandria Ocasio-Cortez, Facebook's Mark Zuckerberg was less than pellucid in his explanation of the scope of the company policy of exempting politicians in political ads from its fact-checking protocol, but his justification for the policy (whatever its scope) was straightforward enough: Zuckerberg said people should be allowed to "see for themselves" what politicians say and draw their own conclusions, rather than have the company decide what they can see.
That rationale sounds broadly similar to the rationale that Justice Kennedy gave for protecting political lies in United States v. Alvarez. Issuing a warning of an Orwellian future, he wrote for the Court: "Our constitutional tradition stands against the idea that we need Oceania’s Ministry of Truth." So too, one might say, we do not want giant companies like Facebook acting as powerful censors, even if the penalty for false speech is only its removal, rather than a rat to the face in room 101.
Yet note that Facebook's policy, if adopted by a state actor, would violate the First Amendment, because it is speaker-based. A long line of SCOTUS cases treats distinctions among speakers as pernicious in the same way that distinctions based on the content or viewpoint of speech. One might think that the Supreme Court misapplied that line of cases when, in Citizens United v. FEC, it condemned a distinction between campaign speech by natural persons and such speech by corporations, but still agree with the general proposition that speaker-based distinctions often do amount to censorship. As the Court put the point in Citizens United: "Quite apart from the purpose or effect of regulating content, ... the Government may commit a constitutional wrong when by law it identifies certain preferred speakers."
That constitutional wrong would very much be committed by a government rule that privileged speech by politicians over speech by ordinary citizens. True, such a rule would not stifle all dissent: People running for office would be permitted to criticize incumbent politicians. However, a rule favoring speech by politicians will naturally tend to squelch dissent and undercut democracy, in which citizens are entitled to tell politicians what they think rather than simply listen quietly.
To be sure, the Facebook policy does not forbid speech by ordinary citizens; it merely disfavors false speech by non-politicians. Yet that distinction seems inadequate. Underwriting Alvarez is the recognition that government ought not regulate the truth or falsity of political speech, regardless of who the speaker is. The case struck down the Stolen Valor Act in general, not just as applied to politicians.
Accordingly, there can be little doubt that Facebook's policy would be unconstitutional if adopted by the government. Does it follow that the policy contravenes basic free speech principles of the sort that Zuckerberg wants to respect voluntarily?
Not necessarily. Some free speech norms for the government do not translate to private actors. For example, a government cannot decide which opinions to stifle and which to amplify, but a newspaper publisher committed to principles of free speech can and will make such judgments in deciding whom to hire as an Opinion writer, what op-eds to publish, and what official editorials to run. Even a newspaper publisher who wants to provide a wide range of opinions will typically deem some opinions on some subjects beyond the pale. The NY Times took a great deal of flak when it hired climate skeptic Bret Stephens to provide "balance." No one at the Times would have thought to hire a Holocaust skeptic, a moon landing skeptic, or a Pizzagate believer to provide balance on those issues, even though the government may not censor Holocaust denial, moon landing skepticism, or Pizzagate conspiracy theories.
Thus, it is at least possible that Facebook could justify its policy on grounds that free speech norms applicable to the government do not apply to it in the same way, even assuming that Facebook wishes to comply with free speech norms.
That's possible, but in practice the claim fails, because Facebook's policy makes little sense. AOC's cross-examination of Zuckerberg showed that he hadn't thought it through very well. And in practice, the policy has already been applied in a way that shows how it favors mainstream views. Taking inspiration from one of AOC's questions for Zuckerberg, Adriel Hampton created a fake ad appearing to show Sen. Lindsey Graham saying "I support the Green New Deal." That didn't fall within the exemption for politicians, because it was speech by Hampton, not Graham, so Hampton registered as a candidate for Governor of California. Facebook still censored Hampton on the ground that he wasn't really a candidate. Maybe that's right, but is Facebook really well positioned to say who's a genuine candidate for office? In 2016 millions of people thought Donald Trump was running for president only as a publicity stunt.
Accordingly, the Facebook policy is misguided and contrary to principles of free speech. What about the Twitter policy? It too would be obviously unconstitutional if adopted by a state actor. The government cannot ban political advertisements. Might there be a justification for Twitter's policy coming from a private actor?
Sure. One justification would be dissociation from controversial speech. Suppose the private owners of a sports arena sell advertising space (as nearly all of them do). They might be perfectly happy running ads for cars, beer, office supplies, and numerous other proprietors of goods and services but unwilling to sell space for political ads. Michael Jordan apparently never said, "Republicans buy sneakers too," but if he had, that would have been a legitimate point about business. The public might not distinguish between the message and the platform, so the owners of the platform have a legitimate interest in not selling political ads--a dissociation interest that is unrelated to censorship.
Consider Games 5 and 7 of the recently completed World Series. When Donald Trump attended Game 5 in Nationals Park, the home fans booed and chanted "lock him up" when the Jumbotron panned to him. When the fans holding a Game 7 Nationals Park watch party of the tv broadcast from Houston saw a Trump ad, they booed again. Good for the Nats fans, I want to say, and so I just did. However, one could see how the Nats owners wouldn't want any fans to feel uncomfortable based on politics at a game, so from their perspective, Trump's attendance and the showing of the ad were unfortunate--although Nationals Park is publicly owned, so the First Amendment applies of its own force, so the owners were somewhat more constrained in their ability to prevent political advertising than they would have been if the Nats played in a privately owned stadium.
In any event, Twitter does not invoke the dissociation interest. Instead, Twitter CEO Jack Dorsey has said that he bans political ads but not Tweets because politicians should earn views on Twitter, not buy them. Is that a good rationale?
Maybe. Dorsey might think that money from the actual marketplace corrupts the Holmesian marketplace of ideas. If so, his view has much in common with the principles behind campaign finance regulations that limit both contributions and spending. To be sure, the Supreme Court has invalidated this sort of rationale in its First Amendment case law, but to repeat, that case law does not apply to Twitter.
Ultimately, whether one thinks the Twitter approach is justified depends on what one thinks of the Court's campaign finance cases. Twitter's approach does not rest on the notion that it is situated very differently from the government with respect to a broad conception of what counts as the corrupting influence of money on politics. Twitter's approach instead rests on the notion that the Supreme Court has misunderstood the First Amendment. Many readers may well agree with Twitter, not the SCOTUS. I know I do.