The Kids Are All Right: Freedom of Expression and New Information Technologies

Mark Tushnet - Harvard Law School

-- Cyberstalking, fake news, revenge porn – today’s social media have seen all these proliferate. And, because all of them are harmful, they have led some to advocate for bans or regulations that would reduce the harm. Those calls have predictably been met with the response that the proposed regulations would impermissibly interfere with freedom of expression.

In "The Kids Are All Right: Freedom of Expression and New Information Technologies", I examine how contemporary U.S. constitutional doctrine would deal with the proposed regulations – or, more precisely, I lay out the analytic framework that U.S. courts would use to determine their constitutionality. The focus on U.S. constitutional law has to be emphasized, for two reasons that point in different directions. First, the long-standing tradition and law of free expression in the United States has been influential around the world on many topics, including the law of sedition and public demonstrations. But, second, U.S. doctrine uses hard-edged categories as the basic unit of analysis whereas most constitutional systems have settled on a more flexible proportionality analysis of free expression issues.

In systems using proportionality the analytic framework to be used to deal with new regulations of new information technologies is simple, though its application in specific cases is not. For example, one common proposal is to make social media platforms like Facebook and Twitter liable for disseminating hate speech or false news stories. Analyzing such proposals means looking carefully at their details: What is the standard for imposing liability? How large would the penalties be for violating the regulations? Some proposals make the platform liable if its operators knew that the posting was fake news, others do so if the operator should have known, still others do so if the operator didn’t have a decent system in place for monitoring postings. Proportionality analysis would ask whether the cost to free expression – the deterrence of the distribution of protected expression – is justified by the reduction in harm, and the outcome might well differ as the details change.

The U.S. approach is interestingly different. It starts with the observation that all forms of expression that elicit regulation cause harm. (If they didn’t, no one would bother regulating them.) The harms come in different forms. Cyberstalking makes its targets fearful and can change the way they behave, revenge porn does all that and also embarrasses its targets, fake news impairs the quality of public discourse and can distort the outcome of elections.

The U.S. approach develops different rules for each type of harm. It sets out the conditions for holding somebody liable for threatening another person, thereby causing fear and potentially changing behavior. The rule is that you can be liable for making a “true threat,” defined either as a threat that a reasonable person would believe had some significant chance of being carried out, or as a threat that the target herself believed might be carried out (I’m stating the rule a bit informally, but adequately for present purposes).

A different doctrine sets out the rules for holding someone liable for making a false statement of fact, knowing it to be false (or recklessly disregarding the possibility that it is false). Here the rule is that you can’t be held liable unless the statement caused or was quite likely to cause what the U.S. Supreme Court called “material” harm: Sending in a false resume for a job application could cause material harm, but saying that you had received a medal for your military service couldn’t.

“The Kids Are All Right” lays out the categories relevant to liability for harms inflicted by people who use today’s social media. It argues that the key feature of those media is that they dramatically reduce the cost of distributing harmful speech and therefore increase the amount of harm that speech causes. But, it argues, in general the increased amount of harmful speech doesn’t require dramatic doctrinal transformations if new regulations are to be held constitutionally permissible.

Some aspects of doctrine might have to be tweaked a bit: You might start with the rule that the person making the threat subjectively intended to place the target in fear, but you might want to tweak it to allow liability when the person making the threat should have known that a reasonable person would have become fearful. Or you might start with a narrow definition of “material harm” for distributing lies but end up with a slightly broader one, for example by treating the effects on the integrity of elections as a material harm. As the article’s title suggests, though, basically U.S. constitutional doctrine is in pretty decent shape to support most of the proposed regulations of new information technologies.

[This isn’t the place to defend either proportionality or the U.S. categorical approach. The choice between them should depend upon institutional details, including the resources available to enforce the regulations, who the regulators actually would be (bureaucrats/civil servants or judges, for example), and more. As a scholar of U.S. constitutional law and comparative constitutional law, I have come to think that on many matters the institutional details in many nations counsel in favor of a hard-edged, U.S.-like approach – but that might simply be a reflection of my own institutional location and intellectual formation.]

30 views0 comments

Recent Posts

See All

Do crowdfunding investors value environmental impact?

Christoph Siemroth - University of Essex; Lars Hornuf - University of Bremen. -- Climate change is increasingly seen as a major societal problem. In our paper, "Do Retail Investors Value Environmental

Collectibles Tokenization & Optimal Security Design

Blair Vorsatz – The University of Chicago Booth School of Business -- Collectibles like art, wine, and classic cars have long interested investors, both for their attractive return profiles and for th