Label AI in political ads
Labels on AI in political ads. Political ads must disclose AI-generated content.
What it does
Political ads using AI-generated or substantially altered images, audio or video would carry a clear notice such as 'This ad was generated or substantially altered using artificial intelligence.' Minor edits like cropping or color would be exempt.
Applies to: Political advertisers and campaigns.
- Political ads must disclose AI-generated content.
Where things stand
Label rules exist in many states, e.g., California AB 2355 (in effect Jan 1, 2025) for committee ads. No federal label requirement; the FEC only says existing fraud rules apply to AI.
Why it costs almost no lead
It applies to campaigns, not labs. Under a day.
Biggest unknown: Whether courts will uphold even pure disclosure labels against First Amendment challenges, since bans have already fallen and label rules have not been squarely tested.
Its effect on p(doom)
Tells voters when a political ad uses AI-generated images, video or voices.
Protects elections from deception. Not aimed at catastrophic risk.
The strongest case that it costs more
Courts have struck down some state versions on free-speech grounds, and labels may not change what voters believe.
The debate
For
- Public Citizen (Ilana Beller) 2026: tracks and pushes state election deepfake laws; petitioned the FEC in 2023
- Sen. Amy Klobuchar and Sen. Lisa Murkowski AI Transparency in Elections Act (2024) requiring disclaimers on substantially AI-generated political ads
- Jessica Rosenworcel (FCC Chair, 2024) Proposed on-air AI disclosure for broadcast political ads, July 2024
Against
- Brendan Carr (FCC Commissioner, now Chair) July 2024 dissent calling the FCC proposal misguided and unlawful
- Christopher Kohls and Hamilton Lincoln Law Institute Won 2025 rulings striking California's election deepfake laws; challenging Minnesota's
- Sean Cooksey (FEC Chair, 2024) Sept 2024: an AI rulemaking would exceed FEC authority
Sources
- Public Citizen tracker: state legislation on deepfakes in elections: State count
- Volokh Conspiracy: California election deepfake law violates First Amendment: Aug 2025 ruling
- Kohls v. Ellison, 8th Cir. 2026: Minnesota law survives preliminary stage
- Federal Register: FEC Artificial Intelligence in Campaign Ads: FEC interpretive rule, no new rule
Rough starting points, not precise forecasts. Lead costs assume China doesn't depend on U.S. models, the case least favorable to safety laws, and count 3 years. On the menu you can change every assumption and put in your own numbers. Last priced 2026-09-26.