OpenAI, Microsoft, and Adobe support California’s AI watermarking bill

OpenAI, Microsoft, and Adobe publicly support AB-3211 The bill would require tech companies to label AI-generated content in California.

The legislation requires that the metadata of photos, videos and audio recordings produced or significantly modified by generative AI systems must be watermarked.

The bill would also require large social media platforms such as Facebook and Instagram to apply “easy-to-understand” labels to uploaded content so users can click to view it. Audio clips and music videos would be labeled visually.

OpenAI expressed its support for the initiative in a letter to California State Assembly member Buffy Wicks, which was published on ReutersAdobe and Microsoft also supported the bill in similar letters. TechCrunch.

“New technologies and standards can help people understand the provenance of content they find online and avoid confusion between human-generated content and realistic AI-generated content,” OpenAI Chief Strategy Officer Jason Kwon wrote in the letter, according to Reuters.

Kwon’s letter also stressed that transparency and attribution requirements are particularly important in an election year, when convincing synthetic content could sway votes. For example, in early August, AI-generated images of Taylor Swift fans supporting Donald Trump circulated on social media.

Adobe provided technical expertise on provenance technology capabilities to assist in the development of AB-3211, said Anne Perkins, Adobe’s director of policy and government relations. “Adobe has always been aligned with the goals of AB-3211 to bring more trust and transparency to our world of AI,” she told TechRepublic.

“Today’s bill provides an effective framework that allows good people to be trusted while protecting people from harmful AI deepfakes, which is especially important during an election year when misinformation is rampant and the stakes are high. We applaud Congressman Weeks for advancing this important legislation that allows AI innovation to flourish while protecting the public from the misuse of this technology.”

AB-3211 has been approved by the state Assembly and Senate Appropriations Committees. It will next go to the full Senate for a vote and then to Gov. Gavin Newsom for a decision on whether to pass it, which is due in late September.

TechRepublic reached out to OpenAI and Microsoft for comment.

look: Apple joins US government’s voluntary commitment on AI safety

Advocacy groups criticized the bill

Microsoft and Adobe are both members of the Content Provenance and Authority Alliance, an industry group working to develop verification standards for AI-generated content. The group helped create C2PA Metadata, a framework for embedding verifiable metadata into digital content used by DALL-E 3.

However, the BSA, an advocacy group for the software industry that includes Microsoft and Adobe, opposed AB-3211 with rhetoric. letter In April this year, the California State Assembly’s Privacy and Consumer Protection Committee received a report that questioned:

• The breadth of content that must be watermarked.
• Request a tool to verify watermarks.
• The notification and reporting system it proposed was “unworkable”.

Lawmakers have since made several changes to the bill that appear to have calmed the companies’ concerns.

California is considering several AI-related bills

AB-3211 is separate from SB-1047. California’s controversial AI bill.

SB-1047 aims to prevent AI models from causing large-scale damage to humans or causing economic loss by imposing strict security requirements on developers. The California Appropriations Committee passed SB-1047 earlier this month, but it still needs to be approved by the state Assembly and Senate before it becomes law.

Much of Silicon Valley, including Open AI, Yuan, Googleand Hug facepublicly disparaged SB-1047. But earlier this week, Elon Musk Bring to X He expressed his feeling that the bill should pass. He posted: “For more than 20 years, I have been an advocate for the regulation of artificial intelligence, just as we regulate any product/technology that poses a potential risk to the public.”

Two former OpenAI employees also wrote letter Newsom criticized his former employer for opposing SB-1047.

“OpenAI even objects to the extremely lax requirements in SB1047, which OpenAI claims are mostly voluntary, casting doubt on the strength of those commitments,” they wrote. OpenAI disagrees with the researchers’ “mischaracterization of its position,” Business Insider

AB-3211 and SB-1047 are only At least 65 AI-related bills Lawmakers plan to introduce the bill in California this year, which aims to balance technological advancement with consumer protection.