First Amendment not a likely impediment to Social Media Transparency Law

Last week, Gov. Gavin Newsom signed Assembly Bill 587, a transparency measure applicable to companies operating the largest social media platforms (Facebook, Twitter, etc.). The law requires these companies to augment the disclosures made in their online terms of service - for example, they must define terms like "hate speech," "extremism" and "disinformation," if they use those terms in regulating content - and to submit a twice-yearly report to the state Attorney General with details and data regarding the platform's approach to identifying and addressing content that violates its rules.

The bill's opponents have suggested that the measure triggers first amendment concerns. This raises the critical question of whether social media companies should be viewed as "speaking" when they censor content posted by users on their platform. The answer will have far-reaching ramifications on how we interact with and regulate (or not) the "modern public square" that is social media.

So far, the 5th and 11th Circuits have weighed in on the constitutionality of social media transparency bills and have reached starkly different conclusions. In NetChoice, LLC et al. v. Att'y Gen. of Fla., 34 F.4th 1196 (11th Cir. 2022), the 11th Circuit considered a Florida law that targeted censorship of speech by or about political candidates, and speech by "journalistic enterprises." The 11th Circuit held that social media companies are private actors with first amendment rights, and that they "speak" by curating and delivering compilations of their users' speech. The court reasoned that a platform's content-moderation decisions were protected exercises of editorial discretion, akin to decisions by a newspaper of what to publish (e.g. Miami Herald Publishing Co. v. Tornillo, 418 U.S. 241 (1974)), or of cable operators to choose the programming they transmit (e.g. Turner Broadcasting System, Inc. v. FCC, 512 U.S. 622 (1994)).

Applying this reasoning, the 11th Circuit applied strict scrutiny to those portions of the Florida law that the court found implicated the platforms' exercise of editorial judgment - such as a prohibition on deplatforming, deprioritizing, or "shadow-banning" candidates - and struck those regulations. However, the court applied intermediate scrutiny to the disclosure provisions of the Florida law and upheld them under the standard articulated in Zauderer v. Off. Of Disciplinary Couns., 471 U.S. 626 (1985), on the grounds that the required disclosures were content-neutral, unlikely to impose a large burden on platforms, and reasonably related to the State's interest in "preventing consumer deception."

Last week, the 5th Circuit took a vastly different approach in upholding a Texas statute that prohibits social media platforms from censoring speech based on the viewpoint of its speaker. In NetChoice, L.L.C. et al. v. Ken Paxton in his official capacity as Attorney General of Texas), 21-51178 (5th Circ., Sep. 16, 2022.), the court "reject[ed] the idea that corporations have a freewheeling [f]irst [a]mendment right to censor what people say," and held that platforms do not engage in "speech," but rather "conduct," when they censor user content. The court disagreed with the 11th Circuit and held that Supreme Court precedent "do[es] not carve out 'editorial discretion' as a special category of [f]irst-[a]mendment-protected expression." Instead, the court asked whether the challenged law either (a) compels the platforms to speak or (b) restricts the platform's own speech - and concluded that the platforms could make neither showing.

The 5th Circuit went further, holding that even if first amendment concerns were triggered, censorship by platforms does not qualify as editorial discretion because - unlike newspapers who "accept reputational and legal responsibility for the content it edits" - platforms expressly disavow responsibility for the content they host. The court buttressed this reasoning by reference to the Communications Decency Act, 47 U.S.C. § 230, which prohibits courts from treating platforms as the "publisher or speaker" of content developed by other users. The 5th Circuit found that this "reflects Congress's judgment that the [p]latforms...are not 'speaking' when they host user-submitted content."

The court also reasoned that while editorial discretion involves "'selection and presentation' of content before that content is hosted, published, or disseminated," the platforms, by contrast, "engage in viewpoint-based censorship with respect to a tiny fraction of the expression they have already disseminated." The 11th Circuit noted that "[s]omething well north of 99%" of posted content does not receive review beyond initial obscenity- or spam-related screening by algorithms, which the court found to undercut a claim of editorial control or judgment.

Also departing from the 5th Circuit's decision, the court held that the Texas legislature permissibly deemed platforms to be common carriers subject to nondiscrimination regulation (though this complex issue deserves to be the subject of its own article).

In a concurring opinion, Judge Edith Jones suggested that platforms - like cable TV companies - do engage in first amendment-protected "speech" by determining which user content to host, but that the Texas disclosure regulations would still pass constitutional muster because (1) the regulations are content neutral; (2) the regulations do not force platforms to modify their own speech, and viewers are unlikely to associate the mandatory hosted speech with that of the platform, and since (3) the platform's selection of content to host controls the flow of information into users' households and could "thus silence the voice of competing speakers with the mere flick of a switch."

As applied here: this author believes that California's Assembly Bill 587 is unlikely to be invalidated on first amendment grounds, because it is limited to the type of disclosure provisions that were upheld in both the 5th and 11th Circuits (albeit for different reasons). Since the law imposes content- and viewpoint-neutral disclosures and does not require or prohibit platforms from engaging in their own speech, it probably passes constitutional muster regardless of whether the 9th Circuit decides that social media giants have first amendment rights to censor. It is possible that speech-related concerns could be triggered if the government chooses to enforce Assembly Bill 587 in a particularly intrusive or burdensome manner, but that is an issue for the litigation that is undoubtedly imminent.