Senior US District Judge Edward Davila recently dealt a blow to the First Amendment rights of social media platforms and to minors’ ability to efficiently view lawful content they seemingly enjoy. He ruled that California may enforce part of a law that, in the absence of “verifiable parental consent,” generally bars platforms from conveying personalized algorithmic feeds of content to minors based on their prior watch history and information “otherwise associated with” minors or their devices.
In rejecting related free-speech challenges filed by Google, Meta, and TikTok, and denying their motions for a preliminary injunction, Davila reasoned they had “not shown that their personalized feeds convey an expressive message or reflect human editorial judgment.” Put differently, the companies failed to show “their First Amendment rights” would be infringed by the statute because, in Davila’s opinion, they lack First Amendment rights when feeding users content they predict users want based on their prior views and characteristics.
The decision is highly problematic for several reasons. First, it allows California to enforce a statute that hinders minors’ ability to easily receive speech that’s fully protected by the First Amendment and—if the companies’ predictive models are accurate—that they also want to receive. That’s important because the US Supreme Court holds that “minors are entitled to a significant measure of First Amendment protection . . . and only in relatively narrow and defined circumstances may government bar public dissemination of protected materials to them.” In short, minors may want to see lawful types of content they’ve previously viewed, but California says it knows better: Minors cannot efficiently receive that speech via personalized-recommendation algorithmic feeds without parental consent.
Second, Davila’s ruling arguably rests on a narrow reading of the US Supreme Court’s 2024 ruling in Moody v. NetChoice that exploits issues the Court there said it wasn’t addressing. In Moody, the Court concluded that “the content-moderation choices reflected in Facebook’s News Feed and YouTube’s homepage” merit First Amendment protection as “editorial choices,” and “that expressive activity includes presenting a curated compilation of speech originally created by others.” Justice Elena Kagan wrote for the majority that “to the extent that social-media platforms create expressive products, they receive the First Amendment’s protection.” A platform seemingly makes editorial choices when its human programmers decide to create algorithms that feed users speech they believe will interest them based on what they’ve previously viewed.
Yet, Kagan added a footnote explaining that the Court does “not deal here with feeds whose algorithms respond solely to how users act online—giving them the content they appear to want, without any regard to independent content standards.” (emphasis added). This reflects something Justice Amy Coney Barrett raised in her concurrence: “[W]hat if a platform’s algorithm just presents automatically to each user whatever the algorithm thinks the user will like—e.g., content similar to posts with which the user previously engaged?” Davila essentially used this opening to underpin his decision in a case he described as testing “the outer limits of when an algorithm-based feed is expressive.”
In so doing, Davila embraced a tenuous distinction between a platform implementing its content-moderation policies about banned content such as hate speech and violent-themed expression, on the one hand, and a platform personalizing algorithmic feeds of third-party content to users based on their “characteristics and history on the platform,” on the other. For Davila, the former may constitute expressive editorial choices about content meriting First Amendment protection per Moody, but the latter doesn’t because it only “reflect[s] back to users their ascertained interests.”
This content-moderation-versus-personalized-feeds dichotomy largely hinges on the presence or absence in a platform’s decision-making of what Davila calls “moral valence” and “moral judgment.” He reasoned that “[c]ontent moderation decisions, implemented through Plaintiffs’ Community Guidelines, carry with them a moral valence; the same is not true for personalization efforts based on users’ data.”
Requiring a speech-affecting decision to carry moral weight before a feed merits First Amendment protection suggests another troublesome distinction: User-agnostic decisions about compiling speech based on normative, subjective value judgments regarding whether “content is bad” under community guidelines deserve constitutional protection, but somehow decisions about prioritizing content that “will be interesting” to users based on “past watch history” do not. Should that be the kind of distinction that determines constitutional protection?
Importantly, Kagan observed in Moody that “Community Standards and Community Guidelines make a wealth of user-agnostic judgments about what kinds of speech, including what viewpoints, are not worthy of promotion. And those judgments show up in Facebook’s and YouTube’s main feeds.” The feeds thus inextricably blend moralistic value choices and give-them-what-they-want decisions; they aren’t based solely on one or the other.
Google, Meta, and TikTok are appealing Davila’s decision to the Ninth Circuit. These are critical cases for both platforms’ and minors’ First Amendment rights that cannot be ignored.