Fake News, Facebook, and Free Speech
by Michael Dorf
My Cornell Government Department colleague Sid Tarrow and I are in the process of putting the finishing touches on a draft of a paper that puts the emergence of fake news in historical perspective, while examining the role that Supreme Court free speech case law and state law play in simultaneously discouraging investigative journalism and encouraging fake news. I'll have more to say about that paper when we are ready to release the draft.
The paper with Prof. Tarrow describes the legal, technological, and cultural forces that give rise to fake news but does not propose a policy response or evaluate others' proposed responses. In this post, I want to offer a few thoughts on one of the responses to fake news: efforts by social media websites, especially Facebook, to identify fake news as fake.
Efforts to label fake news fake raise potential free speech objections. They don't raise First Amendment objections, obviously, because Facebook is not the government, and so the First Amendment doesn't apply to it. However, free speech liberals (like yours truly) have reason to care about the effects of actions by private actors on free speech. For example, there are particular institutions--such as private colleges and universities--that rightly adopt First Amendment or First Amendment-like rules and standards to govern themselves.
More broadly, at least since the Progressive Era, many liberals (e.g., Louis Brandeis) have worried that large private firms can exercise power over the lives of individuals that is roughly analogous to the power that government exercises over them and that therefore constitutional norms ought to be extended to such firms. Typically, the extension comes by way of statute. We might think of anti-discrimination law as an example of this phenomenon--extending something like constitutional equal protection norms to private actors to whom the Constitution does not apply of its own force.
One can quibble at the border about which firms ought to be subject to quasi-constitutional norms. Debates about whether and when religiously scrupled business owners ought to be exempt from various public norms raise such line-drawing questions. Seen through this lens, the individual proprietor running a photography studio or bakery has a better claim to an exemption than does the billionaire family running a giant chain of retail stores--because the latter exercises power over its employees much more analogous to the power exercised by governments. Ironically, the pair of cases just linked denies an exemption to the tiny local business while granting one to the giant retailer.
Where Facebook falls for purposes of the application of public norms is a harder question than it might at first appear. Of course, Facebook is enormous. Many, many more people use Facebook than are employed by Hobby Lobby. However, use of Facebook is entirely voluntary. Facebook users are not vulnerable to Facebook in the way that Hobby Lobby employees or employees of other large firms are vulnerable to their employers.
Moreover, the application of free speech norms by the government to private firms would raise First Amendment questions, because the firms do not merely have an influence on the free speech of their users; they are speakers themselves. Thus, in 1974, in Miami Herald v. Tornillo, the Supreme Court invalidated a state "right of reply" statute on the ground that it limited the editorial discretion of a private newspaper. To be sure, Miami Herald was a departure from the earlier ruling in Red Lion v. FCC, which upheld the so-called "fairness doctrine" as applied to a radio station on the ground that the public airwaves are scarce and thus subject to regulation. But Red Lion has been vigorously criticized on the ground that the the public airwaves are no more scarce than the resources on which newspapers depend and that, in any event, scarcity justifies government licensing but not necessarily content regulation. Red Lion has not been much followed in the years since Miami Herald.
In any event, no one is seriously proposing that the government require Facebook to combat fake news. The question du jour is what steps, if any, Facebook should voluntarily undertake to combat fake news. Facebook is acting in much the way that private colleges and universities act when they voluntarily undertake to promote their own values by adopting--or in certain circumstances, not adopting--First Amendment rules and standards.
In particular, it appears that Facebook is experimenting with a system of labeling fake news fake, by having groups including "Snopes, PolitiFact, The Associated Press, FactCheck.org and ABC News" designate particular stories as "disputed."
This effort seems unlikely to be very effective because some of the very forces that give rise to fake news--ideological polarization, distrust of "elite" authority, and various psychological phenomena such as confirmation bias and motivated reasoning--will beset Facebook's attempt to combat fake news. Consider the fact that Snopes is itself the victim of fake news by right-wingers who claim that Snopes and other fact-checking websites are propagators of left-wing fake news. Indeed, it was probably inevitable that not long after the term "fake news" entered the vernacular, right-wingers would accuse the hated and supposedly left-leaning "mainstream media" of being the real bastion of fake news. In the current environment, a designation of "disputed" from one of the fact-checking watchdogs could come to be regarded by right-wingers as a badge of validation.
Another risk is that right-leaning Facebook users could come to regard Facebook itself as part of the vast left-wing fact-checking conspiracy and migrate away from Facebook to some other social media site. This is obviously a threat to Facebook's bottom line, but it is also potentially bad for public discourse. Already, Facebook users tend to self-segregate into political echo chambers, but there is at least some crossover. People's social networks are sufficiently complex that they usually include others who are not ideologically in lockstep. Those endless Facebook threads in which you argue with your Trump-supporting friends of friends are bad for your mental health but probably good for democracy, insofar as they expose all participants to views they don't generally seek out.
Suppose, however, that Facebook execs conclude that, all things considered, having fact checkers label particular stories "disputed" would reduce the prevalence of fake news--not just on Facebook but in general. Might there still be reasons for Facebook not to proceed with the plan?
One way to think about the question is to ask whether the move, if undertaken by the government, would violate the First Amendment. Suppose that Facebook were government-owned. Would it violate the First Amendment for GovernmentFacebook to label various stories "disputed"?
The answer might depend on how this were accomplished. If the "disputed" label were somehow electronically attached to the stories themselves, that could be seen as compelled speech in violation of cases like Wooley v. Maynard and West Virginia State Board of Ed. v. Barnette. The government has some latitude to compel commercial speech (think about food labeling requirements), but much of what would end up bearing the "disputed" label would be political and other non-commercial speech.
Still, maybe the government could require the application of the "disputed" label without running afoul of the First Amendment. The best precedent for that conclusion is Meese v. Keene, in which the Supreme Court upheld a federal law that required that films and other materials made by foreign governments be labeled "propaganda" when shown in the U.S. The particular films were documentaries on acid rain and nuclear war made by the National Film Board of Canada. The majority opinion relied on the fact that the federal statute defined "propaganda" to refer to materials produced by foreign governments, even if true and of high quality, rather than carrying a pejorative connotation--even though the film's distributor argued that people would predictably infer the pejorative meaning. Based on Meese v. Keene, one might conclude that a government label of "disputed"--if that term were properly defined--did not infringe the free speech rights of the people to whose speech it were applied.
However, even if that's what Meese v. Keene implies, I would urge Facebook to disavow it. Meese v. Keene itself is arguably wrong (as Justice Blackmun argued in a powerful dissent, joined by Justices Brennan and Marshall). And in any event, when a private firm like Facebook decides to adopt First Amendment principles as its own, it need not adopt them in lockstep with Supreme Court jurisprudence. Whether or not Meese v. Keene is rightly decided, it surely violates the spirit of the First Amendment for the government to be labeling some films as propaganda.
Now one might think that there is a crucial difference between the propaganda label in Meese v. Keene and the proposed "disputed" label: Facebook would only apply the "disputed" label to stories that had been judged fake news by independent third party fact checkers, whereas the government labeled even truthful speech propaganda.
That is a vital distinction, of course, but it is important to think about the big picture. I am persuaded by my colleague Steve Shiffrin that the most fundamental aspect of freedom of speech should be protection for the right to dissent. A government or giant social media company that has the power to label fake news "disputed" would be a tempting target for the powerful to co-opt for the purpose of stamping out dissent.
To give an example with which I (as a vegan) have some familiarity, consider the well documented arguments that dairy milk is harmful to human health (e.g., here and here), to say nothing of its impact on the natural environment and the cows and calves exploited and killed in the dairy industry. And yet, our government promotes dairy production and consumption. Such promotion could soon include a federal crackdown on speech. As reported in a recent press release decrying "fake milk," a bipartisan group of 25 members of Congress wrote to the FDA to demand that plant-based milks be barred from using the word "milk." Their letter referred to plant-based milks as "artificial" and "unable to match the nutritional makeup of the product they mimic." Which may be true, but only because the makers of the plant-based milk are understandably not trying to match the harmful nutritional makeup of dairy milk.
Can there be any doubt that the same dairy industry forces that lead members of Congress to call for FDA action to protect consumers from "fake milk" would call on Facebook to protect readers from learning the truth about dairy milk by labeling the materials identifying its true health, environmental, and moral costs as "fake news"? How many other powerful industries and interests would want the "disputed" label attached to information and viewpoints that they regard as dangerous?
As we enter the Trump era, protecting dissent will be more important then ever. Because efforts to label fake news fake pose a potential threat to dissent, they will probably do more harm than good. Fake news may be the price of free speech.
My Cornell Government Department colleague Sid Tarrow and I are in the process of putting the finishing touches on a draft of a paper that puts the emergence of fake news in historical perspective, while examining the role that Supreme Court free speech case law and state law play in simultaneously discouraging investigative journalism and encouraging fake news. I'll have more to say about that paper when we are ready to release the draft.
The paper with Prof. Tarrow describes the legal, technological, and cultural forces that give rise to fake news but does not propose a policy response or evaluate others' proposed responses. In this post, I want to offer a few thoughts on one of the responses to fake news: efforts by social media websites, especially Facebook, to identify fake news as fake.
Efforts to label fake news fake raise potential free speech objections. They don't raise First Amendment objections, obviously, because Facebook is not the government, and so the First Amendment doesn't apply to it. However, free speech liberals (like yours truly) have reason to care about the effects of actions by private actors on free speech. For example, there are particular institutions--such as private colleges and universities--that rightly adopt First Amendment or First Amendment-like rules and standards to govern themselves.
More broadly, at least since the Progressive Era, many liberals (e.g., Louis Brandeis) have worried that large private firms can exercise power over the lives of individuals that is roughly analogous to the power that government exercises over them and that therefore constitutional norms ought to be extended to such firms. Typically, the extension comes by way of statute. We might think of anti-discrimination law as an example of this phenomenon--extending something like constitutional equal protection norms to private actors to whom the Constitution does not apply of its own force.
One can quibble at the border about which firms ought to be subject to quasi-constitutional norms. Debates about whether and when religiously scrupled business owners ought to be exempt from various public norms raise such line-drawing questions. Seen through this lens, the individual proprietor running a photography studio or bakery has a better claim to an exemption than does the billionaire family running a giant chain of retail stores--because the latter exercises power over its employees much more analogous to the power exercised by governments. Ironically, the pair of cases just linked denies an exemption to the tiny local business while granting one to the giant retailer.
Where Facebook falls for purposes of the application of public norms is a harder question than it might at first appear. Of course, Facebook is enormous. Many, many more people use Facebook than are employed by Hobby Lobby. However, use of Facebook is entirely voluntary. Facebook users are not vulnerable to Facebook in the way that Hobby Lobby employees or employees of other large firms are vulnerable to their employers.
Moreover, the application of free speech norms by the government to private firms would raise First Amendment questions, because the firms do not merely have an influence on the free speech of their users; they are speakers themselves. Thus, in 1974, in Miami Herald v. Tornillo, the Supreme Court invalidated a state "right of reply" statute on the ground that it limited the editorial discretion of a private newspaper. To be sure, Miami Herald was a departure from the earlier ruling in Red Lion v. FCC, which upheld the so-called "fairness doctrine" as applied to a radio station on the ground that the public airwaves are scarce and thus subject to regulation. But Red Lion has been vigorously criticized on the ground that the the public airwaves are no more scarce than the resources on which newspapers depend and that, in any event, scarcity justifies government licensing but not necessarily content regulation. Red Lion has not been much followed in the years since Miami Herald.
In any event, no one is seriously proposing that the government require Facebook to combat fake news. The question du jour is what steps, if any, Facebook should voluntarily undertake to combat fake news. Facebook is acting in much the way that private colleges and universities act when they voluntarily undertake to promote their own values by adopting--or in certain circumstances, not adopting--First Amendment rules and standards.
In particular, it appears that Facebook is experimenting with a system of labeling fake news fake, by having groups including "Snopes, PolitiFact, The Associated Press, FactCheck.org and ABC News" designate particular stories as "disputed."
This effort seems unlikely to be very effective because some of the very forces that give rise to fake news--ideological polarization, distrust of "elite" authority, and various psychological phenomena such as confirmation bias and motivated reasoning--will beset Facebook's attempt to combat fake news. Consider the fact that Snopes is itself the victim of fake news by right-wingers who claim that Snopes and other fact-checking websites are propagators of left-wing fake news. Indeed, it was probably inevitable that not long after the term "fake news" entered the vernacular, right-wingers would accuse the hated and supposedly left-leaning "mainstream media" of being the real bastion of fake news. In the current environment, a designation of "disputed" from one of the fact-checking watchdogs could come to be regarded by right-wingers as a badge of validation.
Another risk is that right-leaning Facebook users could come to regard Facebook itself as part of the vast left-wing fact-checking conspiracy and migrate away from Facebook to some other social media site. This is obviously a threat to Facebook's bottom line, but it is also potentially bad for public discourse. Already, Facebook users tend to self-segregate into political echo chambers, but there is at least some crossover. People's social networks are sufficiently complex that they usually include others who are not ideologically in lockstep. Those endless Facebook threads in which you argue with your Trump-supporting friends of friends are bad for your mental health but probably good for democracy, insofar as they expose all participants to views they don't generally seek out.
Suppose, however, that Facebook execs conclude that, all things considered, having fact checkers label particular stories "disputed" would reduce the prevalence of fake news--not just on Facebook but in general. Might there still be reasons for Facebook not to proceed with the plan?
One way to think about the question is to ask whether the move, if undertaken by the government, would violate the First Amendment. Suppose that Facebook were government-owned. Would it violate the First Amendment for GovernmentFacebook to label various stories "disputed"?
The answer might depend on how this were accomplished. If the "disputed" label were somehow electronically attached to the stories themselves, that could be seen as compelled speech in violation of cases like Wooley v. Maynard and West Virginia State Board of Ed. v. Barnette. The government has some latitude to compel commercial speech (think about food labeling requirements), but much of what would end up bearing the "disputed" label would be political and other non-commercial speech.
Still, maybe the government could require the application of the "disputed" label without running afoul of the First Amendment. The best precedent for that conclusion is Meese v. Keene, in which the Supreme Court upheld a federal law that required that films and other materials made by foreign governments be labeled "propaganda" when shown in the U.S. The particular films were documentaries on acid rain and nuclear war made by the National Film Board of Canada. The majority opinion relied on the fact that the federal statute defined "propaganda" to refer to materials produced by foreign governments, even if true and of high quality, rather than carrying a pejorative connotation--even though the film's distributor argued that people would predictably infer the pejorative meaning. Based on Meese v. Keene, one might conclude that a government label of "disputed"--if that term were properly defined--did not infringe the free speech rights of the people to whose speech it were applied.
However, even if that's what Meese v. Keene implies, I would urge Facebook to disavow it. Meese v. Keene itself is arguably wrong (as Justice Blackmun argued in a powerful dissent, joined by Justices Brennan and Marshall). And in any event, when a private firm like Facebook decides to adopt First Amendment principles as its own, it need not adopt them in lockstep with Supreme Court jurisprudence. Whether or not Meese v. Keene is rightly decided, it surely violates the spirit of the First Amendment for the government to be labeling some films as propaganda.
Now one might think that there is a crucial difference between the propaganda label in Meese v. Keene and the proposed "disputed" label: Facebook would only apply the "disputed" label to stories that had been judged fake news by independent third party fact checkers, whereas the government labeled even truthful speech propaganda.
That is a vital distinction, of course, but it is important to think about the big picture. I am persuaded by my colleague Steve Shiffrin that the most fundamental aspect of freedom of speech should be protection for the right to dissent. A government or giant social media company that has the power to label fake news "disputed" would be a tempting target for the powerful to co-opt for the purpose of stamping out dissent.
To give an example with which I (as a vegan) have some familiarity, consider the well documented arguments that dairy milk is harmful to human health (e.g., here and here), to say nothing of its impact on the natural environment and the cows and calves exploited and killed in the dairy industry. And yet, our government promotes dairy production and consumption. Such promotion could soon include a federal crackdown on speech. As reported in a recent press release decrying "fake milk," a bipartisan group of 25 members of Congress wrote to the FDA to demand that plant-based milks be barred from using the word "milk." Their letter referred to plant-based milks as "artificial" and "unable to match the nutritional makeup of the product they mimic." Which may be true, but only because the makers of the plant-based milk are understandably not trying to match the harmful nutritional makeup of dairy milk.
Can there be any doubt that the same dairy industry forces that lead members of Congress to call for FDA action to protect consumers from "fake milk" would call on Facebook to protect readers from learning the truth about dairy milk by labeling the materials identifying its true health, environmental, and moral costs as "fake news"? How many other powerful industries and interests would want the "disputed" label attached to information and viewpoints that they regard as dangerous?
As we enter the Trump era, protecting dissent will be more important then ever. Because efforts to label fake news fake pose a potential threat to dissent, they will probably do more harm than good. Fake news may be the price of free speech.