On August 28, the artificial intelligence (AI) safety bill SB 1047 passed the California State Assembly and now awaits a final decision from Governor Gavin Newsom. The bill has sparked mixed reactions within the tech industry.
AI Legislation Enactment
Senate Bill SB 1047 was authored by Democratic Senator Scott Wiener from San Francisco. Wiener argues that legislation must be established before further advancements in AI become uncontrollable. The bill mandates safety testing for the majority of advanced AI models, either those costing over $100 million to develop or those requiring significant computing power to operate. It will require developers in California to provide outlines for a “kill switch” for such models if they become unmanageable and involve third-party auditors in the safety audits of the models. Developers could also face lawsuits from the state attorney general in the event of ongoing threats from models like AI grid takeovers.
SB 1047 Pushback
While Wiener insists that the bill is crucial to prevent irreversible societal damage from AI systems, not everyone agrees. The bill has faced considerable pushback from key players in the tech industry and state politicians, arguing that it will hinder innovation. Speaker Emerita Nancy Pelosi commented on the bill, calling it “more harmful than helpful” in California’s pursuit to lead in AI development. Neil Chilson, head of AI policy at the Abundance Institute, warned that the bill may affect smaller companies and models, despite being aimed at high-cost and high-caliber models. Tech companies like OpenAI have also expressed disagreement, claiming that the bill will hinder growth, though Wiener disputes these claims. Google and Meta Platforms raised their concerns in a letter to Governor Newsom. Nevertheless, Amazon’s Anthropic supports the bill, stating that its benefits “likely outweigh the costs,” although questions remain.
Other Bills
SB 1047 is not the only AI bill under discussion in California. SB 1220, which bans the use of AI in call centers providing welfare and health services, has also raised questions. However, AB 3211, requiring watermarks on AI-generated content, has gained support from tech companies like OpenAI and Microsoft. Billionaire Elon Musk, developing his model Grok, has expressed support for sweeping AI safety regulations.
The SB 1047 bill has divided opinions within the tech industry and among politicians. Governor Gavin Newsom is expected to make a decision on the matter by September 30. Cointelegraph has reached out to legal experts and AI developers in California for a deeper understanding of the situation.
Comments