California’s push to regulate artificial intelligence (AI) development has progressed significantly with the Safe and Secure Innovation for Frontier Artificial Intelligence Models Act, or SB 1047. State Senator Scott Wiener spearheaded the legislation, emphasizing the urgent need to balance innovation with safety. The act has garnered diverse reactions from stakeholders, including AI companies, lawmakers, and industry experts.
The bill, co-authored by Senators Richard Roth, Susan Rubio, and Henry Stern, aims to set comprehensive safety standards for the development and deployment of advanced AI models exceeding $100 million in training costs. Significant amendments have been integrated into the bill following feedback from AI firms such as Anthropic, as well as federal lawmakers from California’s Bay Area, including Representatives Zoe Lofgren and Nancy Pelosi.
Wiener acknowledged that while the amendments do not encompass all of Anthropic’s suggestions, the core concerns have been addressed. The revised bill has removed criminal penalties, replaced with civil ones, along with the deletion of the proposed Frontier Model Division. “We accepted a number of very reasonable amendments proposed, and I believe we’ve addressed the core concerns expressed by Anthropic and many others in the industry,” said Wiener.
Prominent AI figures like Geoffrey Hinton and Yoshua Bengio have publicly supported SB 1047. Hinton emphasized the sensible balance the legislation maintains, stating, “Forty years ago when I was training the first version of the AI algorithms behind tools like ChatGPT, no one – including myself – would have predicted how far AI would progress. Powerful AI systems bring incredible promise, but the risks are also very real and should be taken extremely seriously.”
Nonetheless, other tech giants remain wary. Companies such as Hugging Face and venture capital entities like Andreessen Horowitz have expressed apprehensions that the bill might stifle innovation and drive firms to relocate outside California. This apprehension is reinforced by comments from Martin Casado of Andreessen Horowitz, who labeled the amendments as “window dressing” that fails to mitigate the bill’s underlying issues.
The legislation has advanced out of the Assembly Appropriations Committee, which underscores its significant budgetary implications. It sets forth several rules to ensure robust safety measures for AI models, including the capacity to completely shut down models that pose catastrophic risks or critical harm to public safety. The act mandates employing an auditor to secure compliance and defines substantial training costs as exceeding $100 million.
SB 1047 faces a pivotal vote in the California Assembly, which must occur by the end of August. The bill’s passage is eagerly monitored nationwide as it could potentially set a precedent for state-level AI regulation. If approved, it will proceed to the state Senate and then to Governor Gavin Newsom for final endorsement.
While the legislation has ignited debate within Silicon Valley and beyond, it reflects a critical attempt to reconcile innovation with safety in the rapidly evolving AI landscape. “California must act to get ahead of the foreseeable risks presented by rapidly advancing AI while also fostering innovation,” Wiener asserted, highlighting the gridlock at the federal level on comparable technology regulation.
News Sources
- California law for AI safety has been sugar-coated to please corporations
- California AI Catastrophe Bill Clears Committee
- California bill to regulate AI advances over tech opposition with some tweaks
- Controversial California AI Law Moves Forward
- California trims AI safety bill amid fears of tech exodus
Assisted by GAI and LLM Technologies
Source: HaystackID
