Invest2 publishers3 min readPublished
California prices a knowing under-16 feed violation at $50,000 per minor from January 2027
California's under-16 feed ban starts on January 1, 2027, and the cheapest way to comply is an age signal built from a birth date the user typed at device setup. The EFF expects platforms to verify every adult instead.
The Investor · Invest desk

What happened
- Newsom signed a 13-bill safety package that from January 1, 2027 bans algorithmic feeds and autoplay for social media users under 16 and requires chatbot crisis protocols when minors show signs of suicidal ideation.
- AB 1709 sets civil penalties of $50,000 per minor for knowing violations and up to $25,000 for negligent ones, enforceable only by the California Attorney General or a local public prosecutor.
- A companion bill, AB 2, creates civil liability for platforms found negligent in harming children, with penalties reaching up to $1 million per child.
- The Digital Age Assurance Act, signed in October 2025, makes operating system providers collect birth dates at device setup and pass app developers a bracketed age signal from January 1, 2027.
- Two AI audit laws signed the day before the package established what techtimes.com calls the nation's first independent third-party AI verification framework.
Compiled by The InvestorSomething wrong?How this is made
Why it matters
- decision Every covered platform now picks between the free self-declared operating system signal and identity checks run on its whole California user base, and the choice is made on liability.
- cost A separate under-16 build, or a California cutoff for that age group, comes out of the same engineering budget as ranking and feed work.
- constraint Because only the Attorney General or a local prosecutor can sue, a platform can size its reserve to the chance of a filing.
- contradiction The EFF reads the regime as pushing verification onto every adult, and the same regime rests on a birth date the user types in, so the invasive reading and the porous reading both follow from the text.
Section 22684 is the operative sentence for anyone costing this out: a covered platform "shall verify the age of a user" before providing restricted features, and the bill creates liability per child every time a minor passes through an age check undetected [14].
There are two ways to satisfy that. The free one is the bracketed age signal that operating system vendors must pass to app developers under the Digital Age Assurance Act [16]. The other is a check the platform runs itself, on everybody, since you cannot sort out the under-16s without also sorting the over-16s. The Electronic Frontier Foundation asked Newsom to veto AB 1709 and called it "well-intentioned, but deeply flawed" [12]. On the group's reading, the statute forces the second option, because it names no minimum verification method and imposes strict liability for errors, so the most invasive verification available lands on every California user regardless of age [15].
The weakness of the free path is that the birth date is self-declared. A 15-year-old who types an adult date at device setup sends an 18-and-older signal to every app on the phone [17]. A platform leaning on that signal is buying its compliance at $25,000 per undetected minor if a court calls the conduct negligent and $50,000 if it calls the conduct knowing [5]. Two hundred misses is $5m or $10m [1]. AB 2 sits on top at up to $1m per child, twenty times AB 1709's knowing-violation figure [6][2].
What caps the exposure is procedural. AB 1709 has no private right of action, so a case has to come from the California Attorney General or a local public prosecutor [5]. A large platform can integrate the free signal, set a reserve, and wait to see whether anyone files; the deterrent then runs on prosecutorial appetite rather than plaintiff volume.
The other cost is product. Platforms that cannot technically separate addictive features from the rest of the service for under-16s have to rebuild the product for minors or shut them out of California entirely [8]. Either way, ranking loses that engineering time. The EFF's substantive objection is that the definition of addictive features is broad enough to prohibit virtually every functional recommendation algorithm for that age group. The group says that will hit LGBTQ+ youth and other marginalised users who depend on online communities hardest [13]. Assemblymember Josh Lowenthal, who wrote AB 1709, called it and its companion liability measure "the most impactful social media protections bills in the nation" [9][7]. Newsom, signing at a children's museum in Marin County, said: "Parents can't compete. Can't compete with the engineering, can't compete with the algorithms." [11][2]
This is the third technology-safety package Newsom has signed in two years, a sequence techtimes.com traces back to the 2022 Age-Appropriate Design Code [19]. The Digital Age Assurance Act gave device vendors about 14 months between signing and activation [3]. AB 1709's platform obligations land on the same date [16][3]. If covered platforms ship just the API integration and 2027 passes with no enforcement action, the reading that this law verifies every adult in California is wrong. The cost was a few weeks of integration work.
What to watch
- Whether the California Attorney General or a local prosecutor files the first AB 1709 action after January 1, 2027, and against which platform.
- Whether the operating system vendors' bracketed age signal ships with anything stronger behind it than the birth date a user types at setup.
- Whether any covered platform bars California under-16s instead of building a separate product for them.