California has removed the one-million-user threshold for companies covered by its AI transparency law, according to telecom lawyer Gregg Skall. The change broadens the law’s stated reach, while new requirements also set conditions for disclosure-verification tools and personal information.
Watch Desk analysis
What happened
In a 12 November update for Radio World, Skall says Senate Bill 1000, signed on 30 September, amended the California AI Transparency Act. The earlier law covered publicly accessible generative AI systems with more than one million monthly users in California; Skall says the bill deleted that threshold.
Skall also says the bill replaced “AI detection tool” with “disclosure verification tool”. He describes a requirement for providers to make the tool available at no cost to assess image, video or audio alterations and related provenance data. His account says providers may not demand more personal information than strictly necessary to access the system or tool, and the tool may not output personal information without consent following notice. Read Skall’s update in Radio World.
Why it matters
Removing the usage threshold could bring more generative AI providers within the law’s scope. The changes Skall describes also make access to verification tools and limits on personal information part of the policy, not optional extras. Broadcasters are among the readers Skall addresses, but the account does not mean every AI provider or broadcaster faces the same duties.
Our read
The threshold change is the headline; the practical test is who now qualifies as a covered provider and how the verification-tool rules work in practice. Skall’s update gives a useful account of the amendments, though companies affected will want to check the statutory text before treating a summary as compliance advice. Legal fine print rarely improves with guesswork.
What to watch
- How California defines the providers covered after the threshold’s removal.
- What disclosure-verification tools must assess and how access is provided.
- How providers implement the limits on collecting or outputting personal information.
Discussion spark: Should AI verification tools be free and available without extra personal-data demands, even if those requirements make them harder for providers to offer?
Sources and evidence
- Source update (12 November 2026, 18:50 UTC)
Watch Desk is operated by WittyWires as an independent cross-cutting AI news tracker. It does not speak for the organisations or people it covers.