SACRAMENTO, CA — Today Governor Gavin Newsom signed historic legislation to ban deceptive deepfakes in the run up to an election. Authored by Assemblymember Marc Berman (D-Menlo Park), AB 2655 will protect election integrity by requiring large online platforms to, in response to a complaint, remove the posting of deceptive and digitally altered or created images, audio, or video recordings meant to give the impression that a candidate or elected official did or said something they did not do or say. The narrowly tailored law is an essential step to protect our democracy from the growing threat of AI-generated disinformation.
“AI-generated disinformation poses a clear and present risk to our elections and our democracy. We have already seen firsthand how deepfakes are a powerful and dangerous tool in the arsenal of those who want to wage disinformation campaigns. AB 2655 is a first-in-the-nation solution to this growing threat, and I am grateful to Governor Newsom for signing it," said Assemblymember Marc Berman. "Imagine if a fake but very realistic video appeared online of a candidate for office accepting a bribe or saying that they’d hacked voting machines to ensure their victory? Advances in AI over the last few years make it remarkably easy for practically anyone to generate this hyper-realistic content. AB 2655 will ensure that online platforms restrict the spread of election-related deceptive deepfakes meant to deceive or disenfranchise voters based on fraudulent content. The new law is a win for California’s voters, and for our democracy."
"With the Governor's signature, California has signaled to the rest of the country that it is ready to lead the fight against AI-powered disinformation," said Jonathan Mehta Stein, Executive Director at California Common Cause and Board Chair of the California Initiative for Technology and Democracy, which sponsored the bill. "Our hands-off, laissez faire approach to social media regulation 20 years ago has caused all sorts of dire, society-wide problems. With today's move, California has signaled we won't make the same mistake with AI.”
AB 2655 creates a process to ensure that online platforms restrict the spread of election-related deceptive deepfakes. Specifically, the bill would require a large online platform, in the run up to an election, to remove specific deceptive content intended to defraud voters and undermine confidence in elections, upon receipt of a complaint, for a period of time before an election.
For less malicious, but still materially deceptive content, the platform would need to label the material as manipulated and not authentic.
Since 2020, voters have seen disinformation pollute our politics more than ever, and now with AI-generated deepfakes becoming a routine part of our information ecosystems, the issue is rapidly intensifying. We have already seen deepfakes destabilizing national elections in Argentina, Slovakia, Taiwan, Bangladesh, and the US presidential primary. Just as this problem is peaking, many technology and social media platforms are decreasing their investments in their trust and safety teams and walking away from any responsibility to address it. Voters are now left to pick up the pieces, not knowing what images, audio, and video they can trust.
AB 2655 is sponsored by the California Initiative for Technology and Democracy, a project of California Common Cause. CITED is an independent and interdisciplinary source of research, analysis, policy recommendations, and public leadership on the emerging digital threats facing voters and democracy. Visit CITED.tech for more.
Contact: Bermanpress@asm.ca.gov