California Gov. Gavin Newsom has signed a groundbreaking law in a law that requires the world’s largest artificial intelligence company to release safety protocols and report serious cases, state lawmakers announced Monday.
Senate Bill 53 shows California’s most important move to maintain its position as a global high-tech hub while regulating Silicon Valley’s rapidly advancing AI industry.
“Using transformative technology like AI, we have a responsibility to support that innovation. We will introduce common sense guardrails,” state Sen. Scott Winner, sponsor of the bill, said in a statement.
The new law represents the success of Wiener’s second attempt to establish safety regulations for AI after Newsom rejected its previous bill, SB 1047, after a furious pushback from the tech industry.
It also comes after a failed attempt by the Trump administration to prevent states from enacting AI regulations under the debate that it would create regulatory disruption with China and slow down US-made innovation.
The new law says that large AI companies must publish safety and security protocols in compiled format to protect their intellectual property.
You should also report critical safety incidents to state officials within 15 days, including model-enabled weapon threats, major cyberattacks, and loss of model control.
The law also establishes protections for employees whistleblowers that reveal evidence of danger or violation.
According to Wiener, California’s approach differs from the European Union’s groundbreaking AI law, and requires private disclosure to government agencies.
Meanwhile, SB 53 requires public disclosure to ensure greater accountability.
Describing it as the world’s first provision, the law requires companies to report cases in which AI systems engage in dangerous, deceptive behavior during testing.
For example, if the AI system is in the effectiveness of controls designed to prevent the construction of Bioweapon, developers should disclose the incident if they substantially increase the catastrophic risk.
The working group behind the law was led by prominent experts, including Stanford University’s FEI-FEI LI, known as the “AI Godmother.”
