Why OpenAI is opposing California's AI safety bill
OpenAI's Chief Strategy Officer, Jason Kwon, has voiced concerns over a proposed AI safety bill in California. In a recent letter, Kwon insisted that AI regulations should be under federal jurisdiction rather than state control. He warned that the new bill could potentially slow technological progress and prompt companies to relocate out of the state.
OpenAI advocates for federally-driven AI policies
Kwon, in his letter, argued for a federally-driven set of AI policies instead of a patchwork of state laws. He believes this approach would promote innovation and position the US to lead the development of global standards. "A federally-driven set of AI policies, rather than a patchwork of state laws, will foster innovation and position the US to lead the development of global standards," Kwon stated in his letter.
Details of the controversial AI safety bill
The bill in question, SB 1047, also known as the 'Safe and Secure Innovation for Frontier Artificial Intelligence Models Act,' was introduced by California State Senator Scott Wiener. The proposed legislation aims to establish standards for developing powerful AI models. It mandates pre-deployment safety testing, whistleblower protections for AI lab employees, and grants legal power to California's Attorney General if AI models cause harm. The bill also proposes a "public cloud computer cluster" named CalCompute.
Senator Wiener defends AI safety bill
In response to Kwon's letter, Senator Scott Wiener defended the proposed bill. He pointed out that the requirements would apply to any company doing business in California, regardless of their headquarters' location. Wiener also noted that OpenAI did not criticize any specific provision of the bill. He concluded by stating, "SB 1047 is a highly reasonable bill that asks large AI labs to do what they've already committed to doing, namely, test their large models for catastrophic safety risk."
Bill awaits final vote amid amendments and opposition
The controversial bill, after facing opposition from politicians like Nancy Pelosi and Zoe Lofgren, companies such as Anthropic, and organizations including California's Chamber of Commerce, has passed out of committee with several amendments. These modifications include replacing criminal penalties for perjury with civil ones and limiting the Attorney General's pre-harm enforcement abilities. The bill is now awaiting its final vote before reaching Governor Gavin Newsom's desk.
OpenAI pauses expansion amid regulatory concerns
Amid concerns about California's regulatory landscape, OpenAI has reportedly put discussions about expanding its San Francisco offices on hold. The company, which has had an office in San Francisco's Mission district for years, recently moved into a new office in the city's Mission Bay region, previously occupied by Uber. This decision comes as the company continues to express concerns over SB 1047.