AI transparency bill moves forward in California with Transparency Coalition backing
Transparency Coalition co-founder Jai Jaisimha joined California State Sen. Josh Becker (D-Menlo Park) to testify on behalf of Becker’s AI Transparency Act (SB 942) yesterday in Sacramento.
Speaking before the California Assembly Privacy and Consumer Protection Committee, Jaisimha said the Act “will help bring much-needed transparency to generative AI outputs, and it will provide the public with critical insight into which content was created in whole or part by generative AI systems.”
SB 942 requires large generative AI providers to embed tools within their system to disclose that certain content was created by AI. It was approved by the California State Senate last month and is now making its way through the Assembly.
Sen. Becker said his bill addressed “the challenges we’ve already seen with deepfakes, in both popular culture and election manipulation, as well as fraudsters exploiting generative AI for scam calls.”
“Clearly,” Becker added, “transparency is needed.”
Jaisimha highlighted some of the key elements in the bill, including:
The inclusion of manifest disclosures embedded into the AI-created image or video
Embedding latent disclosures into the data associated with the AI-generated content
Meaningful penalties for violation of the Act.
“We don’t think these are heavy-handed requirements, nor do we think they’ll inhibit small business innovation,” Jaisimha told the committee. “The technology behind it is proven. Even OpenAI is supportive of C2PA and has actually started using it for DALL-E, one of their models.”
C2PA is a technical standard that allows developers and others to embed metadata in media to verify its origin and related information.
Following Becker and Jaisimha’s testimony, the Assembly Privacy and Consumer Protection Committee vote 8-0 to approve the bill. It now moves on to the Assembly Judiciary Committee. For more analysis of the bill, see our previous post, Three California AI bills are worth supporting as they face key tests this month.
In other AI news yesterday, the Privacy and Consumer Protection committee also approved SB 1047, the Safe and Secure Innovation for Frontier Artificial Intelligence Models Act authored by State Sen. Scott Weiner. That bill also moves on to the Assembly Judiciary Committee.