OpenAI Back, California Bill Requiring 'Digital Watermark' On AI Generated Content
OpenAI, the mastermind in behind ChatGPT, has announced its support for a California bill aimed at regulating AI-generated content. The proposed legislation seeks to mandate tech companies to clearly label AI-created content, encompassing a wide range of materials from harmless memes to deceptive deepfakes designed to disseminate misinformation about political candidates. By backing this bill, OpenAI strives to mitigate the risks of misinformation and manipulation, promoting a safer and more responsible AI ecosystem.
AB 3211, a California bill requiring labels on AI-generated content, has been overshadowed by SB 1047, which mandates safety testing on certain AI models. AB 3211 aims to prevent misinformation and manipulation by ensuring transparency in AI-generated content.
The California bill, AB 3211, has already cleared several hurdles, passing the state Assembly unanimously and the Senate Appropriations Committee. If it passes the full Senate by August 31, it will head to Governor Gavin Newsom for signature or veto by September 30, potentially becoming a law.
The tech industry, including OpenAI and its backer Microsoft, has pushed back against the bill, expressing concerns and opposition to its requirements
California lawmakers introduced a sweeping 65 bills this session to address various AI-related concerns, including ensuring unbiased algorithmic decisions and preventing the exploitation of deceased individuals’ intellectual property by AI. Although the legislative effort aimed to establish comprehensive regulations, many of the proposed bills have been rejected, narrowing down the field to a smaller set of remaining proposals still under consideration by lawmakers.
OpenAI advocates for transparency and accountability in AI-generated content, emphasizing the importance of provenance measures like watermarking, particularly during election seasons. In a letter to California Assembly member Buffy Wicks, OpenAI expressed support for regulatory efforts to ensure clarity and trust in AI-driven content, aligning with Wicks’ bill promoting transparency in AI-generated materials.
Many countries, representing a third of the world’s population, hold elections this year, experts worry about the impact of AI-generated content on the voting process. This concern is heightened by its notable presence in recent elections, such as in Indonesia.
As per Reuters report, OpenAI’s Chief Strategy Officer, Jason Kwon, stated in a letter that new technology and standards can help people identify the source of online content and distinguish between human-created and realistic AI-generated content, reducing confusion.
(Nazwin Basheer has produced this article as part of a Timeline Internship Programme)
