US children online safety age verification

Sacramento's Warning Label War: Why California's AB 56 Will Lose in Court — and Should

California's social media warning-label law for minors faces a serious First Amendment challenge — and a familiar legal fate.

AB 56 in Context People of Internet Research · US Jan 2026 AB 56 effective date California's warning-label and def… AB 2273 Prior CA law enjoined The Age-Appropriate Design Code Ac… 1995 Anonymous speech precedent McIntyre v. Ohio established const… 39M+ Adults affected by ID checks California's population subject to… peopleofinternet.com

Key Takeaways

On January 1, 2026, California's AB 56 — signed by Governor Gavin Newsom in late 2025 — took effect, requiring social media platforms to display state-scripted mental-health warning labels to minors and to impose default time limits and notification restrictions on users under 18. Four months in, the law is already where most of California's youth-internet legislation ends up: in federal court. NetChoice, the trade group that has now defeated a long series of state online-speech mandates, has sued. In May 2026, the Electronic Frontier Foundation filed an amicus brief warning that AB 56's warning-label mandate and its de facto age-verification regime amount to compelled speech and a threat to anonymous internet access for every Californian, not just teenagers.

The instinct behind AB 56 is understandable. The U.S. Surgeon General's 2023 advisory Social Media and Youth Mental Health urged platforms, parents, and policymakers to take adolescent well-being seriously, and called for clearer disclosures. Lawmakers in Sacramento took that as license to mandate the disclosures themselves. But good intentions do not unlock the First Amendment, and California's track record on this exact question is unflattering.

A familiar legal cliff

California has spent the last two years losing variations of this fight. In NetChoice v. Bonta, the Ninth Circuit in 2024 affirmed an injunction against core provisions of the California Age-Appropriate Design Code Act (AB 2273), holding that the state's compelled "data protection impact assessments" — which required platforms to opine on whether their content was harmful to children — likely violated the First Amendment. The U.S. Supreme Court's 2024 decision in Moody v. NetChoice reinforced the foundational point: a platform's choices about what speech to host, amplify, or restrict are themselves protected editorial activity.

AB 56 sits squarely in that minefield. A government-scripted message that platforms must show to a specific audience is the textbook definition of compelled speech under cases stretching from Wooley v. Maynard to NIFLA v. Becerra. The state will argue the labels are purely factual disclosures akin to cigarette warnings under Zauderer, but the science here is materially different from tobacco. The Surgeon General's own 2023 advisory acknowledged that the research linking social media use to adolescent mental-health harms is mixed, correlational, and highly dependent on what teens are actually doing online. A categorical state warning that flattens that nuance is closer to ideological messaging than to a factual nutrition label.

The age-verification problem nobody voted for

The more consequential design flaw in AB 56 is one most Californians never debated. To deliver different defaults and different warnings to under-18 users, platforms must know who is under 18. That means age verification — and at scale, age verification means identifying everyone. As EFF's May 2026 EFFector explains, laws like AB 56 "in the name of online safety" effectively force platforms to build mandatory identity infrastructure that adults are then conscripted into using.

Anonymous and pseudonymous speech is not a loophole. The Supreme Court protected it explicitly in McIntyre v. Ohio Elections Commission (1995) and again in Talley v. California (1960). Whistleblowers, abuse survivors, dissidents from authoritarian states using U.S. platforms, LGBTQ+ teens in unsupportive households — these users do not have a spare government ID to upload to TikTok or Meta. AB 56 does not require ID checks on its face, but its under-18 carve-outs are unenforceable without them. That is how the bill becomes, in EFF's framing, a social-media regime for everyone.

What proportionate regulation actually looks like

The pro-innovation position is not that adolescent harms are imaginary. They are real for some kids, in some contexts, on some products. The position is that California has chosen the least effective and most constitutionally fragile tool available.

The likely outcome

Based on the Ninth Circuit's reasoning in NetChoice v. Bonta, the warning-label provisions of AB 56 are unlikely to survive strict or even intermediate scrutiny. The default-restriction provisions for minors are on stronger footing and may be severable. The most useful thing the federal courts can do is exactly what they have been doing: distinguish between regulating platform conduct (data minimization, dark-pattern bans, default settings) and regulating platform speech (compelled warnings, content classification). California keeps blurring that line. The courts keep redrawing it.

Californians who genuinely want safer online experiences for teenagers should want AB 56 struck down quickly, so the legislature is forced to draft something that will actually take effect.

Sources & Citations

  1. EFF: California's Terrible, No Good, Very Bad Social Media Ban (May 2026)
  2. EFF + 18 orgs on UK online-harm roots (May 2026)
  3. MediaNama: OpenAI backs KOSA (May 2026)
  4. U.S. Surgeon General Advisory: Social Media and Youth Mental Health (2023)
  5. NetChoice litigation tracker