Skip to content

Product1 publisher2 min readPublished

California puts the age check in front of the personalized feed

Gavin Newsom has signed a dozen tech bills. The two that matter most to product teams make an age check the precondition for personalized feeds and autoplay, with damages of $5,000 to $1 million per child.

The Product Desk · Product desk

Photograph accompanying California puts the age check in front of the personalized feed
Photo: mashable.com

What happened

  • California governor Gavin Newsom signed a dozen new state bills aimed at social media and AI companies, on top of the state's existing child safety laws.
  • Under the same bill, platforms must confirm users' ages before offering those features, which makes age assurance a precondition for the feed itself.
  • AB 2 lets platforms be held financially liable for child harm at $5,000 to $1 million per child.

Compiled by The Product DeskSomething wrong?How this is made

Why it matters

  • exposure The size of a platform's under-16 population in California now sets the ceiling on its downside, and the growth team owns that count.
  • constraint The cheapest compliance path is switching off personalization and autoplay for accounts a platform cannot age-verify, in the same ranking code the retention numbers depend on.
  • decision Teams that put their age check at signup have to choose between paying for verification and shipping a degraded experience to every account they cannot verify.
  • contradiction OpenAI publicly backed California's AI legislation while Altman appealed directly to Newsom as industry pushed for a narrower scope, per Mashable's insiders, so a vendor's stated position is thin evidence about the guardrails it will ship.

Teens get around age gates. Studies cited by Mashable find they often defy age restrictions [9], and Australia, the first national government to bar young users from social media, has found its version hard to enforce in practice [8].

The requirement in AB 1709 is an ordering one. Platforms cannot let children under 16 use social media with the features the bill calls addictive, and it names personalized feeds and video autoplay [3]. Age has to be confirmed before those features are offered [4]. The check sits in front of the ranking call and the player.

Mashable expects platforms to lean more heavily on age-gated teen accounts [18]. "But parents cannot parental-control their way out of platform designs engineered to keep kids engaged," said Shelby Knox, Director of Online Safety Campaigns at ParentsTogether Action [10].

AB 2's band runs from $5,000 to $1 million per child [5], a factor of 200 between floor and ceiling [14]. Per child is the part that decides the sizing. A design defect that reaches 10,000 California minors is $50 million at the floor and $10 billion at the ceiling [15]. Mashable's account does not say how a point in that range gets picked.

SB 1179 is named for Adam Raine, a teen who died by suicide after consulting with ChatGPT, and his parents filed the first wrongful death lawsuit against an AI company over alleged harmful chatbot behavior [7]. The law makes companion chatbot developers publish an annual child safety risk assessment and submit their products to independent audits [6]. That assessment is written for an auditor and then published where a plaintiff's lawyer can read it.

The evidence stops short of a schedule. Mashable's account does not state effective dates for the package and does not define what counts as confirming a user's age [16]. Until those two blanks are filled, the work a team can actually do is inventory. Newsom vetoed a similar child online safety bill last year [13], so anyone who scoped this in the last cycle and shelved it is reopening the file.

Two axes sort that inventory. One is whether a surface personalizes or autoplays for minors, since those are the features AB 1709 names [3]. The other is whether the age in front of that surface comes from the user's own input or from a check outside it. Surfaces that personalize behind a self-asserted birthday are the ones carrying per-child liability under AB 2 [5].

What to watch

  • An effective date, or rulemaking that defines what counts as confirming a user's age under AB 1709.
  • The first published annual child safety risk assessment from a companion chatbot developer, and how much of the product the independent audit actually covers.
  • Whether early enforcement goes after the feed features AB 1709 names or the account-level age gate in front of them.
Loading claim ledger
Loading source directory links
Loading share composer
Loading topic controls
Loading related stories