The California Assembly has approved a controversial artificial intelligence safety bill that would require companies to guarantee that their technology will not cause significant harm.
SB 1047 has received both criticism and praise from prominent figures across the country. The bill will be sent back to the state Senate for a confirmation vote and amendments before it reaches Governor Gavin Newsom. The bill passed the state Assembly with 41 votes in favor and 9 against.
SB 1047 requires companies developing powerful AI models to use “reasonable care” to ensure their technology does not cause “significant harm,” such as multiple casualties or property damage exceeding $500 million. The bill, introduced by state Sen. Scott Wiener and passed the state Senate in May, requires companies to take preventative measures, such as implementing a kill switch that can shut off their technology at any time. It also requires models to undergo third-party testing to ensure they minimize significant risks.
Additionally, the bill would create whistleblower protections for employees of AI companies who want to share safety concerns, and companies that don’t comply could be sued by the California Attorney General.
“Innovation and safety can go hand in hand, and California is leading the way,” Wiener said in a statement. “With today’s vote, the Legislature has taken a truly historic step to proactively protect the public interest as exciting new technologies advance.” He also called the bill a “light-touch, common-sense measure” that codifies safety efforts that many large companies already have in place.
The bill has faced fierce opposition over the past few months from many large technology companies, startups, and venture capitalists, who argue that the bill amounts to government overreach focused on a technology that is still in its early stages and could stifle innovation in states. Last week, OpenAI publicly voiced its opposition to the bill, arguing that such policies should be implemented at the federal level, not the state level.
Politicians including former House Speaker Nancy Pelosi, Rep. Ro Khanna and San Francisco Mayor London Breed have also spoken out against the bill, reflecting tech industry concerns that it could hinder California’s leadership in AI innovation.
But the bill has also garnered support from some of the biggest names in AI in its final days before the vote. On Tuesday, Elon Musk unexpectedly voiced his support, though he said it was “a tough call and it will upset some people.” Anthropic, a rival to OpenAI known for its safety-mindedness, also offered cautious support for the bill, saying last week that “the benefits likely outweigh the costs” and that its provisions could be implemented. Anthropic doubled down on its support for the bill after Wiener implemented some of the amendments it proposed.
Wiener defended the bill against criticism, emphasizing that the bill’s provisions only apply to companies that have spent more than $100 million training large-scale models or $10 million fine-tuning them, not most small startups. He also said he supports federal legislation but believes Congress has been historically reluctant to regulate technology and that in the absence of national action, it is California’s responsibility to lead the way.
The fight over SB 1047 will continue now that the bill has been sent to Governor Newsom’s desk. Critics of SB 1047 include OpenAI, tech incubator Y Combinator, and venture capital firm Andreessen Horowitz, all of which have registered lobbyists working on the bill.
The new special issue features a Wall Street legend getting a dramatic makeover, tales of cryptocurrency shenanigans and badly behaved poultry royalty, and more.
Read the story.
