Just published at 2 J. Free Speech L. 157 (2022), as part of the “Non-Governmental Restrictions on Free Speech” symposium; here’s the Abstract:
First Amendment protection of “editorial discretion,” “editorial control,” and “editorial judgment” has a relatively short Supreme Court history. First used in the 1970s, these terms refer to the power that broadcasters, cable systems, and newspapers retain to make decisions about their content within regulatory regimes. Editorial decision-making is an action that editors perform on others’ speech—which sometimes expresses and conveys an editor’s own message, other times not. When, as in Miami Herald Publishing Co. v. Tornillo, editorial decisions express and convey editors’ ideas, they receive First Amendment protection. But, as in FCC v. Midwest Video Corp, when they do not—or the regulatory regime at issue allows editors to express their own views, First Amendment protection is limited or non-existent. If all editorial decision-making were to receive full constitutional protection, long-established common carrier law as well as mandatory carriage of political advertisement and PEG programming would be rendered unconstitutional.
Internet platform content moderation decisions are mostly non-expressive editorial decisions. Platforms’ content moderation decisions in toto do not convey a message because, lacking a fixed expression, they are never communicated in toto to anyone—as the platforms do not publicize their decisions. Unlike First Amendment-protected editorial decisions, such as inclusion in an op-ed page, a cable system channel line-up, or a parade, an audience cannot read a list of content-moderation decisions and thereby comprehend an expressed message.
Second, platforms’ editorial decisions express little individually. For instance, shadow banning, by which a platform renders a user’s posts invisible to all but that user, cannot convey a message because no one knows, except the platform, that it is happening. Invisible editorial decisions are not intrinsically expressive of a corporate policy because, as the Court recognized in Rumsfeld v. FAIR, such conduct only communicates by reference to other speech.
Last, just as telephone companies do not express their customers’ conversations, transmitting a message, in the context of a communications network, does not express a social media platform’s own editorial decisions or speech, a position that platforms themselves have maintained vociferously in countless section 230 cases. Because content moderation decisions are largely unexpressive, social media laws such as Texas’s H.B. 20 are consistent with the First Amendment.