A California judge just called your feed a mirror, not free speech

On 5 August, Senior US District Judge Edward Davila refused to block key parts of California’s Protecting Our Kids from Social Media Addiction Act. The three companies had asked for a preliminary injunction against the law, known as SB 976. Davila denied it, ruling they were unlikely to win their First Amendment challenge.

The contested provisions bar platforms from sending minors “addictive feeds”, meaning personalised streams built from a user’s past behaviour, without verifiable parental consent. They cap a minor’s default feed at one hour a day and switch off personalisation by default. By 2027, operators must also assess whether a user is a minor, part of a wider age-check push.

Why the free-speech argument failed

The platforms leaned on Moody v. NetChoice, the 2024 Supreme Court ruling that a platform’s content-moderation choices are protected expression. Davila drew a sharp line. Moderation can be expressive, he agreed. Personalisation is a different thing.

A personalised feed, he wrote, is “not an ‘expressive’ message; it is merely a mirror that reflects back to users their own perceived interests”. The companies were “merely relying on predictive modelling” to guess what would keep a user scrolling. That, the judge held, is number-crunching, not editorial judgment.

He used a plain example. A platform does not push cat videos because its boardroom thinks cats are good, but because an algorithm predicts the user will watch. Content moderation carries a “moral valence”, he wrote, so deciding to bar self-harm content is a value judgment. Ranking by engagement is not.

What it means, and what comes next

The ruling matters beyond California. It hands other states a template for regulating engagement-driven feeds without tripping the First Amendment. The trick is separating what a platform chooses to allow from what its algorithm chooses to amplify. At the June hearing, the state cited a girl who lost consciousness after a “blackout challenge” video reached her feed.

It arrives as Meta faces a wider reckoning over teen harm, from a wave of addiction suits to a New Mexico court that just ordered it to pay $567m. TikTok, with more than 170 million US users, is settling its own teen cases. The pattern is a legal system testing whether platform speech defences hold as harms to minors pile up.

Davila was careful about the limits. This is a preliminary ruling, and fuller evidence “could reveal that provisions not enjoined now are unconstitutional”. For now, California can enforce the law. The platforms had no immediate comment. The deeper question, whether a feed is speech or a machine, is only beginning.