On Wednesday, October 2, a federal judge in the United States approved a preliminary injunction blocking the implementation of AB2839, the deepfakes law signed by California Governor Newsom just two weeks ago. This law allows individuals to file civil lawsuits seeking compensation for deceptive images, audio, and video materials generated using AI technology for election purposes.
AB2839, titled “Elections: Deceptive Media in Advertisements,” was proposed by Democratic State Assemblywoman Gail Pellerin from the 28th district. Pellerin stated that deceptive information generated by AI could influence voter behavior and undermine confidence in elections by various entities such as conspiracy theorists, foreign governments, internet trolls, and even political campaigns themselves.
The law stipulates that during the 120 days leading up to an election and the 60 days following it, candidates, committees, or election officials may seek injunctions or other remedies to prohibit the dissemination of deceptive AI-generated or altered content in advertisements and other election materials. They are also authorized to pursue special damages in civil suits.
AB2839 passed in both the Senate and Assembly with significant margins and was set to be effective until January 1, 2027, with provisions for extensions or new legislation thereafter.
Senior District Judge John A. Mendez expressed concerns about deepfakes but emphasized that such fears should not allow legislators to overturn the protections of the First Amendment, which safeguard criticism, parody, and satire – vital components of democratic debate.
Mendez wrote that much of AB2839 seemed to act as a blunt instrument rather than a surgical tool, stifling important freedoms and uninhibited exchange of ideas necessary for American democracy.
Theodore Frank, a senior attorney at the Hamilton Lincoln Law Institute (HLII), welcomed the court’s ruling, asserting that new technologies should not alter the principles protected by the First Amendment.
The First Amendment Coalition’s Legal Director, David Loy, argued that the government should not create new speech categories beyond the protections of the First Amendment and emphasized the need for a comprehensive legal framework to determine fraud or defamation.
The coalition raised constitutional concerns regarding AB2839, stating that allowing lawsuits against anyone disseminating related content, regardless of personal harm to the plaintiff, could inappropriately embroil courts in political disputes.
Just two days after Newsom signed the law on September 19, Christopher Kohls, a YouTube content creator represented by HLII, filed a lawsuit challenging the new legislation. Kohls had circulated a parody video featuring AI-modified voice of Democratic presidential candidate and current Vice President Kamala Harris.
According to HLII’s website, Newsom responded to the video by stating that manipulating voices in “advertisements” like this should be illegal. Newsom’s post indicated his intention to ensure its illegality through upcoming legislation.
Shortly after Newsom signed the bill, Elon Musk, the CEO of social media platform X, shared the AI video on X, stating that declaring such parody videos illegal went against the U.S. Constitution. Free speech advocates also argued that California’s new law infringed on First Amendment rights.
Pellerin declined to comment on the lawsuit, emphasizing that the video was digitally altered for parody purposes, not malicious intent.
Apart from AB2839, Newsom signed two more bills related to election advertising and content dissemination – AB2355 and AB2655 – in September. These laws aim to address the deceptive use of AI-generated and altered content in campaign activities, remove such content from major online platforms, increase accountability, and inform voters.
AB2655, proposed by Democratic State Assemblyman Marc Berman from the 23rd district, requires major online platforms to remove or label deceptive, digitally altered/created content related to elections within specified periods, along with providing mechanisms to report such content. It also authorizes various officials to issue injunctions against platforms.
AB2355, presented by Democratic State Assemblywoman Wendy Carrillo from the 52nd district, mandates disclosure of whether election advertising materials have been AI-generated or significantly altered, allowing the Fair Political Practices Commission to seek injunctions or other remedies.
These three new California laws are considered among the strictest in the U.S. around similar legislation. Newsom emphasized the importance of upholding election integrity in a press release, highlighting the necessity to prevent AI from undermining public trust through false information during the current politically tense climate. These laws aim to protect individuals from unauthorized replication of their voices or images and combat the use of deepfakes in political advertising.
Newsom’s spokesperson Izzy Gardon stated that these laws safeguard democracy and preserve free speech, believing that the courts will support California’s regulation of such dangerous and misleading deepfake works, emphasizing that satire remains prevalent in the state even for those who miss the humor.