U.S. AI Safety Institute Signs Agreements for Early Model Access With Anthropic and OpenAI
The U.S. AI Safety Institute (AISI) has signed agreements with Anthropic and OpenAI, that will give the agency early access to major new models from each company prior to and following their public release, giving AISI an opportunity to provide feedback to Anthropic and OpenAI on potential safety improvements to their models. The government also plans to conduct collaborative research with the companies on how to evaluate capabilities and safety risks.
California State Legislature Passes Bill Regulating Large-Scale AI Systems
The California legislature has approved the AI Accountability and Safety Act (SB 1047), which now California Governor Gavin Newsom must decide whether to sign into law. The bill seeks to regulate large-scale AI systems by requiring safety tests and allowing the California Attorney General to sue if these technologies cause serious harm. The congressional members argue that the bill could negatively impact California’s economy and innovation.
Read more about why the California bill would undercut more sensible federal efforts.