Start the day here

Law — Social Media — Free Speech

Judge Davila Denied Meta's Claim That Feeds Are Speech

Judge Edward J. Davila of the Northern District of California refused on August 5 to block three provisions of Senate Bill 976, California's Protecting Our Kids from Social Media Addiction Act. Meta, TikTok, Google, and YouTube had asked for a preliminary injunction. They walked out without one.

The platforms argued that compiling and ordering third-party posts for each user is itself expression, the digital cousin of a newspaper editor choosing the front page. Davila rejected that frame. Ranking systems, the order says, are "designed to evaluate users, not messages." The resulting feed "is merely a mirror that reflects back to users their own perceived interests," a passive receptacle that Justice Alito's Moody concurrence would call a dumb pipe.

6 min read
A smartphone home screen on a wood table showing Facebook, Instagram, Messenger, and YouTube icons

What California can now enforce

Section 27001 bars addictive personalized feeds for minors without verifiable parental consent (or without actual knowledge the user is a minor). Section 27002(b)(2) sets a one-hour daily default unless a verified parent raises it. Section 27002(b)(4) lets parents force a default feed that does not rank on the child's behavioral signals. Age-assurance regulations are due by January 1, 2027. Coverage reporting on the 22-page order is here.

The line that matters

Davila split two functions the companies insisted were one. Filtering illegal or banned material from a candidate pool is expressive: community rules carry a moral valence written by people. What comes next is different. Meta's engineers described prediction models such as p(like) and p(video completion). TikTok scores videos for interest. YouTube optimizes for long-term satisfaction. Those systems guess what will keep a thumb moving. They do not announce a viewpoint.

The hearing transcript does the rest of the work. Platforms say they show content because it is interesting or trustworthy, Davila noted, "but [Plaintiffs are] not filling in that blank." Filling in the blank is editorial judgment. Setting an engagement objective leaves the blank empty.

Fallback theories failed too. Users' right to receive information still leaves search and affirmative requests intact. Claims that the whole feed is protected speech collapsed once personalization was held non-expressive. YouTube's vagueness attack on "verified parent" lost to ordinary dictionary meaning.

The platforms had supplied the declarations the Ninth Circuit said NetChoice could not. Company-specific engineering detail was supposed to prove expression. It proved the opposite: a rules gate, then a score. That architecture is the real news. Separate the expressive filter from the engagement ranker, and the ranker becomes ordinary regulatory terrain instead of strict-scrutiny speech.

Other desks are already treating design as design. The European Commission has preliminary DSA findings against TikTok and Meta over addictive defaults. Massachusetts' top court has said Section 230 does not swallow state design claims. Davila's order is preliminary, and fuller evidence could reopen the constitutional fight. For now, California can enforce the three sections, and every state watching has a sentence it can steal: the feed is a mirror. Mirrors do not get the First Amendment.

Related stories

Two hands write with a mechanical pen across a stack of white papers on a dark deskAcute Social Issues

Judge Lin Vacated Anthropic's Supply-Chain Risk Label

Today

A dark wooden gavel with a brass band resting on its sound blockAcute Social Issues

Is AI Consciousness Just a Liability Shield?

Today

Red candlestick chart sloping downward on a dark trading screenAcute Social Issues

DeepMind Measured How Gemini Steers Beliefs and Money

Today

Letters

0

No letters yet.

Write a letter