Golden State Today

Legislation addressing the threat of AI-powered disinformation to elections has advanced in the California state legislature. Despite opposition from the technology industry, a package of bills sponsored by the California Initiative for Technology and Democracy (CITED) passed the Assembly floor, with some measures receiving bipartisan support.

“This is a massive, bipartisan step forward in tackling the dangers that unregulated AI poses to our voters and to our democracy, while still respecting the First Amendment and innovation,” said Jonathan Mehta Stein, executive director of California Common Cause. “Where Congress won’t act, California and CITED will lead the way in ensuring that our elections and our democracy are protected from new digital threats.”

The legislative package, introduced in March, aims to regulate disinformation amplified by AI and social media. Central to the package are two bills that prohibit deepfakes of candidates, elected officials, and election officials during a reasonable period before Election Day and shortly after during vote counting. For instance, under these bills, a robocall impersonating "Joe Biden" urging Democrats not to vote would be illegal.

Since 2020, voters have increasingly encountered sophisticated disinformation due to advancements in generative AI technology. A November 2023 poll by Berkeley IGS showed that 84% of California voters are concerned about digital threats to elections, with 73% believing state government has a “responsibility” to act. This concern spans across all demographics.

“There have always been bad actors who try to wreak havoc on elections,” said Assemblymember Gail Pellerin. “But now the widespread access to generative AI is making that type of damage so much more sophisticated and easier to disseminate.”

“AB 2655 will ensure that online platforms restrict the spread of election-related deceptive deepfakes meant to prevent voters from voting or deceive them based on fraudulent content,” stated Assemblymember Marc Berman. He added that deepfakes could significantly disrupt democracy by falsely attributing speech or actions to individuals.

CITED’s work draws on research from experts in tech, law, public policy, civil rights, civic engagement, and academia. The group was launched in 2023 by California Common Cause to provide independent policy leadership on emerging digital threats.

If enacted into law, CITED’s bill package could set a precedent for reforms in other states and at the national level.