The Federalist Society’s recent National Lawyers Convention featured a panel titled, “Private Control Over Public Discussion.” The panel, which focused on proposals to regulate the content moderation decisions made by prominent social media firms, featured a presentation by Georgetown Law Professor and Cato Institute Senior Fellow Randy Barnett. Barnett presented an argument for social media regulation grounded in public accommodation law that I have not addressed yet. While among the most interesting arguments in favor of requiring social media firms to host more content, it fails thanks to a few conceptual flaws.

Those who have been following the policy debates surrounding social media content moderation will be familiar with a number of arguments for policies that would force social media companies to host more content. Many of these arguments rest on weak foundations. For example, PragerU argued that YouTube is a “company town” like Chickasaw, Alabama. In Marsh v. Alabama (1946) the Supreme Court held that First Amendment protections applied in Chickasaw, a town owned by a private company. The 9th Circuit correctly noted in its PragerU v. Google ruling that the Google-owned YouTube is nothing like a company town. Unlike the company town under consideration in Marsh, YouTube does not run the municipal functions of a town.

Others, such as Barnett’s co-panelist Michigan State University Prof. Adam Candeub, argue that social media firms should be subject to common carriage regulation. I have written about this argument before and debated Candeub last September. I am not convinced that common carriage regulation for social media firms is appropriate. Such an approach rests on a misunderstanding of the Internet infrastructure and how social media companies present themselves to the public. Those interested in reading more of my thoughts on common carriage regulation should read these posts.

Barnett’s argument does not rest on common carriage regulation or company town status. Rather, Barnett embraces an understanding of public accommodation law and applies it to prominent social media companies. As best as I can tell, Prof. Barnett has not presented this article in writing anywhere. What I will discuss in the next few comments is based only on his comments at the recent National Lawyers Convention.

Barnett does not outline specific policy recommendations. The talk is best understood as an exploration of an idea: the application of public accommodation law to social media companies. Barnett deploys an analogy using the civil rights movement of the 1960s. During the civil rights movement many protestors and activists used their freedom of movement, as guaranteed under the 14th Amendment, to travel. Many non-government institutions such as hotels, motels, and restaurants discriminated against African Americans and their allies, giving rise to the famous Green Book.

Barnett argues that institutions within civil society need not be “private non government” or “public government.” A third category, developed over hundreds of years of common law, is available: “public non-government.” Institutions in this category are “affected by a public interest.” Hotels, motels, gas stations, and restaurants fall into this category. They might be private, but they also operate along critical infrastructure (i.e. roads) and offer vital services. Similarly, Barnett argues, social media firms operate on critical infrastructure (i.e. the Internet) and offer vital services. Accordingly, perhaps social media firms should be subject to viewpoint anti-discrimination policies similar to the racial anti-discrimination policies found in accommodations laws.

There are a number of reasons why I think applying public accommodation laws to social media firms would be inappropriate. As I will argue below, the analogy between social media companies and motels is weak. Even if the analogy was conceptually sound, applying public accommodation laws to social media firms would result in a social media environment that very few people anywhere on the political spectrum would find amenable to debate, advocacy, education, or journalism.

In his comments at the Federalist Society conference Barnett said the following:

I suggest that to the extent that a private company has created a forum for the public to communicate their ideas, such a company is limited to barring speech that the Supreme Court has found to be unprotected from government restriction. If a governmentally provided public forum cannot restrict such speech then neither can a non-governmentally provided public forum.

The first problem with this framing is that of definitions. Plenty of online spaces can be considered a “forum for the public to communicate their ideas.” Social media platforms such as those run by Twitter, YouTube, Facebook, and Reddit come to mind, but the definition could as easily apply to Wikipedia, Rotten Tomatoes, Telegram, Discord, Twitch, WhatsApp, Yelp, and many, many others. Even comments sections belonging to traditional institutions such as newspapers could be considered “for the public to communicate their ideas.” “Social media” is much in the news these days, but it is a difficult term to define. If Barnett’s argument is a solution to the perceived problem of anti-conservative content moderation in “Big Tech” firms it will require a more specific definition in order to avoid over-inclusion.



Perhaps the most notable problem with Barnett’s argument is that social media companies do not hold themselves out to the public in the same way that motels and restaurants do. Prominent social media companies do not hold themselves out to the public as venues for all legal speech. Each of the most prominent social media firms list policies that describe legal speech they choose to prohibit on their platforms, including discrimination against some political speech.

Nor are social media firms necessary to civic society in the way motels and restaurants were during the civil rights era. Civil rights activists driving across the Jim Crow South had nowhere to turn if a gas station owner in a remote location refused to serve them. The same is not true for Facebook and YouTube. Those seeking to speak online have a plethora of options to choose from. If you are booted from Twitter you can speak online elsewhere with ease.

That is not to dismiss the inconveniences of being booted from a popular online platform such as Twitter or Facebook and losing out on beneficial network effects. But with an Internet ecosystem that includes alternatives to “Big Tech” platforms such inconveniences do not justify public accommodation regulation.

It might be the case that the Internet is as critical a piece of infrastructure as roads are, but Google, Facebook, and Twitter are not the Internet. An argument in favor of applying accommodation laws to web hosts such as GoDaddy, Cloudflare, and AWS would be stronger as the services they provide are more akin to critical infrastructure than Facebook and Google. Such arguments may be particularly appealing in the wake of complaints over AWS suspending service to Parler.

If social media firms were prohibited from barring legal speech they would lose much of their utility. Harassment, spam, and other categories of content most online speakers and readers detest would be difficult or impossible to address. Debates and discussions that would otherwise be valuable would run the risk turning into a morass of irrelevant and objectionable content. Some might argue that such content is already a feature of contemporary online debates. Yet if you think that there is plenty of objectionable content on Facebook, Twitter, and YouTube with their content moderation rules in place imagine how much worse online speech will become when these firms must treat all legal speech equally and are forbidden from engaging in viewpoint discrimination. A glimpse at Facebook’s “Dangerous Individuals and Organizations List” should give you some idea.

Those discussing social media regulation should keep in mind the influence of “Big Tech” social media firms. Some have taken to describing Facebook and other prominent social media sites as operating a “public square.” But this is to exaggerate their influence. Most Americans are not on Twitter. Facebook and YouTube are more popular than Twitter, yet less than a quarter of Americans claim that they “often” get their news from social media. Facebook is losing popularity among teens and young adults. We should recall that it was not that long ago that commentators were discussing MySpace’s monopoly.

Barnett’s argument is a welcome change to a debate that is unfortunately often characterized by misunderstandings of law and content moderation policies. Barnett’s argument is safely grounded in an understanding of the law. Its flaw is the assumption that a “public non-government” category, which is applied to motels, restaurants, and gas stations, is appropriate for social media platforms. Many are upset with how Facebook, Twitter, and YouTube have treated some third party speech. Fortunately, there are other venues for online speech available, and those concerned with ideological bias in Silicon Valley remain free to speak elsewhere.