A judge halts California's new AI law in a case related to a deepfake of Kamala Harris.

A federal judge halted the implementation of a new law on artificial intelligence in California, just a few days after Governor Gavin Newsom signed it. The law, known as AB 2839, was created with the purpose of holding accountable those who distribute deepfakes on social media, particularly those that alter the image of political candidates and may mislead voters. Newsom had suggested that this regulation could be applied to compel Elon Musk to remove a deepfake of Vice President Kamala Harris that he had posted.

U.S. District Judge John Mendez ruled in favor of Christopher Kohls, who had originally posted the deepfake, arguing that the law is unconstitutional. Kohls contends that his content is a form of satire protected by the First Amendment. In his ruling, Mendez issued a preliminary injunction that temporarily prevents the California Attorney General from enforcing the law against Kohls or anyone else, except for audio messages governed by AB 2839.

Judge Mendez questioned the vagueness of the law, indicating that any altered digital content could be misinterpreted as harmful. He emphasized that although the state has a legitimate interest in maintaining electoral integrity, the breadth and ambiguity of the law could result in a violation of free speech rights. He noted that existing laws, such as those regarding privacy rights or defamation, could already protect public figures affected by altered representations of their image.

The halt of this law could have significant implications for the landscape of free speech in the digital realm, especially considering that AB 2839 is just one of eighteen new artificial intelligence laws recently signed by Newsom. While it remains to be seen whether the suspension will turn into a permanent block, it is clear that this outcome represents a victory for free speech advocates on the platform X.