On 4 November 2026, Meta, Google (including YouTube) and TikTok each sued California Attorney General Rob Bonta in federal court to block Assembly Bill 1709, according to MediaNama. Governor Newsom signed the law in September. It takes effect on 1 January 2027. It bars personalised feeds and autoplay for users under 16 even when a parent consents. It also carries penalties of up to $50,000 per affected user. The companies argue the law violates the First Amendment and conflicts with Section 230.
The strongest case for the law
The case for AB 1709 deserves a fair statement. Engagement-optimised feeds and autoplay are design choices. They are built to extend a session, and a 13-year-old is less equipped than an adult to resist them. Parental consent rules have a weak record because consent screens are easy to click through and hard to enforce. A flat prohibition removes the burden from families and puts it on the party that designed the feature. California has also argued before that this is regulation of product design, not of speech.
Why the law is still a poor fit
The law's flaws are in its construction, not its goal.
First, it overrides parents. The law applies even where a parent approves, so the state replaces a family's judgment with its own. Google's reported objection is that the law takes decision-making away from parents and interferes with YouTube's ability to recommend age-appropriate material. A recommendation system is also how a child gets pointed to educational video instead of whatever is most viral. Banning personalisation does not guarantee better content. It may only mean less curated content.
Second, the enforcement mechanism is severe. Platforms must take reasonable measures to keep minors away from the features and must delete the accounts of under-16 users who keep receiving them. A penalty of up to $50,000 per affected user can grow to an existential figure for any platform with a large teen base. That pushes companies toward blunt age verification and over-removal. Those costs fall on adults and on privacy, and smaller services will struggle to absorb them.
Third, the law is hard to separate from content. A rule about which items a service chooses to show a user is a rule about what the service says to that user.
What the courts have already said
Two precedents frame the case, and they point in different directions.
In Moody v. NetChoice, decided 1 July 2024, the Supreme Court vacated and remanded the Florida and Texas platform laws. It did so because neither lower court had properly analysed the facial First Amendment challenges. The Court noted that the Eleventh Circuit had treated restrictions on content moderation as triggering scrutiny under cases protecting "editorial discretion", while the Fifth Circuit had called moderation "not speech" at all. The Court did not settle every feed question, but it required a platform-by-platform and function-by-function analysis.
The Ninth Circuit applied that approach in NetChoice v. Bonta, filed 9 September 2025, which concerned California's earlier SB 976. The panel held that some personalised feed algorithms may be expressive, but that this inquiry is fact-intensive. It also held that NetChoice had failed to show unconstitutional applications substantially outweighing constitutional ones. The district court's finding that NetChoice lacked associational standing was also upheld, because the case lacked information about individual members. The panel did enjoin SB 976's like-count requirement as content-based.
This history shows why the three companies are now suing in their own names. A trade association struggled to prove an as-applied case on behalf of its members. Each platform can now put its own recommendation system into the record and argue that its feed reflects editorial choices.
The Section 230 argument is the weaker one
The companies also say AB 1709 conflicts with Section 230, which provides that no provider of an interactive computer service shall be treated as the publisher or speaker of information provided by another. That argument has intuitive force when a state penalty depends on how third-party content is surfaced. But a law that regulates a design feature, and not particular posts, is a harder target for a statutory immunity defence. The First Amendment claim is the main event.
A national patchwork is forming
The California fight is part of a wider shift. In mid-October 2026, two Eleventh Circuit opinions allowed Florida and Georgia to begin enforcing their laws restricting young people's social media access, according to the Electronic Frontier Foundation. Platforms now face different rules by state, enforced under different tests. That uncertainty is itself a cost to innovation, and it argues for a consistent federal approach.
A proportionate path
The stronger policy options sit between a ban and inaction:
- Require safer defaults for teens, such as time limits and autoplay off, that a parent can change.
- Mandate transparency about how recommendation systems affect minors, so evidence can accumulate.
- Target specific harms with narrow rules rather than a category of design.
- Fund digital-literacy and parental-tool adoption, and measure results.
A court that allows AB 1709 would give states room to dictate how any platform arranges its content. A court that strikes it down would not end the debate. It would only push lawmakers toward narrower rules that are easier to defend and more likely to work. Either way, the ruling will turn on facts about how these feeds actually operate, which is exactly what Moody and the Ninth Circuit asked litigants to prove.