Judge halts California's election deepfake ban
A federal judge just temporarily blocked California from enforcing its new AI deepfake law, AB 2839, which targeted users distributing deceptive political content. The ruling followed a lawsuit over an AI parody of Kamala Harris that Elon Musk reposted, with the court deciding the statute is too broad and threatens political satire. Why it matters: The injunction stops the state's crackdown on election deepfakes just weeks before voters cast their ballots.
California tried to ban election deepfakes, but a federal judge just pumped the brakes.
Less than two weeks after Governor Gavin Newsom signed assembly bill AB 2839 into law, U.S. District Judge John Mendez issued a preliminary injunction blocking the state from enforcing it. The move prevents the state's attorney general from applying the law against virtually anyone, leaving only provisions around standalone audio messages untouched for now.
The drama started when Elon Musk reposted an AI-generated parody video of Vice President Kamala Harris, originally created by X user Christopher Kohls. Newsom publicly implied the new bill could force Musk to take it down. Kohls hit back by suing the state the day after the signing, arguing the content was protected satire under the First Amendment.
Targeting posters instead of platforms
What made AB 2839 unique was its focus: instead of holding tech platforms liable, it went straight after the people distributing the deepfakes. If an individual posted content resembling a political candidate knowing it could confuse voters, California judges could force them to delete it or face cash fines.
Judge Mendez wasn't having it. In his ruling, he wrote that the statute relies on subjective terms that effectively drag vast amounts of political and protected speech into danger.
He noted that YouTube videos and X posts are simply today’s political cartoons and newspaper ads. While fears over digital media manipulation are understandable, Mendez said they do not give lawmakers unbridled license to bulldoze over critique, parody, and satire, pointing out that existing defamation and copyright laws already offer legal recourse.
Why it matters
This is a massive win for online creators and Musk’s free speech corner on X. It sends a clear message that slapping an AI label onto parody does not strip away constitutional protections. If you make political memes or satirical videos, you will not have to worry about California court orders knocking on your door anytime soon.
It is also a reality check for aggressive AI regulation. Newsom recently signed 18 different AI-related bills, but AB 2839 was crafted with such broad strokes that it practically invited a constitutional challenge. We think attempting to police individual posters rather than platforms was a legally fragile gamble that backfired immediately.
Because this is a preliminary injunction, the legal battle will drag on, but the law is effectively sidelined for the upcoming November election. Lawmakers everywhere now face a tougher challenge: you cannot simply outlaw bad-faith media without accidentally outlawing satire.
Sources (1)
- techcrunch.com(opens in a new tab)Originaltechcrunch.com
Topic trackerCalifornia Election Deepfake Law Legal Challenge
- Judge halts California's election deepfake banThis story
- Elon Musk reposts Kamala Harris deepfake, testing X's own rules