State Senator Josh Becker of Menlo Park worked with EU lawmakers to develop legislation requiring embedded data on AI-generated images and audio that can be verified and tracked, and the first phase took effect in both places Saturday.
California's new AI transparency requirements officially took effect Saturday alongside the European Union's AI Act, marking the first phase of a coordinated effort to make AI-generated content easier to identify, as KQED reports. The law, which passed in 2024, requires generative AI companies to embed difficult-to-remove “provenance” data into AI-generated photos, videos, and audio, including whether the content was made using AI, which system was used, and the date it was made. That information can then be verified using content authentication tools such as those developed by Coalition for Content Provenance and Authenticity.
State Senator Josh Becker (D-Menlo Park), who co-authored the legislation with Assemblymembers Buffy Wicks (D-Oakland) and Rick Chavez Zbur (D-Los Angeles), said the goal is to help combat election misinformation, scams, and abusive deepfakes by giving the public a way to distinguish authentic content from fabricated media.
“While AI has some benefits, we've seen AI content used for election misinformation, for scams and abusive deepfakes,” Becker said, speaking to KQED.
Becker told KPIX in an interview Monday that California lawmakers worked closely with the European Union while developing the legislation so both measures could take effect simultaneously and help establish a common international framework, adding that Washington state has already adopted similar legislation.
“This is now a global standard really because of my bill,” Becker said, speaking to KPIX.
The law's next phases will roll out over the coming years. Becker explained in a reel on social media Monday that beginning January 1, large social media platforms operating in California must provide users with a way to view embedded provenance information attached to AI-generated content. Then, beginning in 2028, phones and cameras manufactured in California will also be required to capture and store provenance information the moment the content is created.
Becker said the measure took nearly three years to implement after it passed in 2024 because lawmakers wanted to ensure the technical standards were practical before requiring companies to comply while also acknowledging that technology is advancing faster than regulation and that California will need to move more quickly on future AI policy.
He said California is uniquely positioned to influence AI regulation because it is home to much of Silicon Valley and the world's fourth-largest economy. Becker said many leaders within the AI industry have also acknowledged the need for guardrails, adding that California's partnership with the European Union gives the law influence well beyond the state's borders.
“I represent a million people, most of Silicon Valley,” he told KPIX. “So I think we have an opportunity but also responsibility to regulate this industry.”
While the measure won't eliminate the need for skepticism online, the intention of the law was to give people better tools. Becker is scheduled to hold a press conference Monday in Sacramento with representatives from the European Union to discuss the new law and its implementation.
Related: Google to Add Clearer Sources In AI Searches, Allow Sites to Opt Out Following UK Ruling
Image: Senator Josh Becker/Instagram