California lawmakers pass AI safety bill SB 53 — but Newsom could still veto

The California Senate gave the final approval early Saturday morning to the main integrity of the Amnesty International Safety law, which defines new transparency requirements for large companies.

It was also described by its author, senator in the state, Scott Winner, SB 53 require [employees] In AI LABS it creates a public cloud to expand access to a Calcompute account. “

The bill now goes to the California Governor Gavin New Love to sign or veto. He did not publicly comment on SB 53, but last year, he objected to a more expansionist safety bill than Wiener, with narrower legislation to be signed targeting issues such as Deepfakes.

At that time, newsom admitted the importance of “protecting the public from the real threats posed by this technology”, but criticized the previous Wiener invoice to apply “strict standards” on large models regardless of whether it was “published in high -risk environments, [involved] Make critical decisions or use sensitive data. “

Winner said that the new draft law was affected by recommendations from a policy committee from the artificial intelligence experts held by Newsom after the veto.

Politico also mentioned that SB 53 has been recently modified so that companies that develop “Frontier” models of artificial intelligence need less than $ 500 million in annual revenue only to reveal high -level safety details, while companies that make more than that will need to provide more detailed reports.

The bill has been opposed by a number of silicon valley companies, VC companies and pressure groups. In a back message to newsom, Openai SB 53 was not specifically mentioned, but he argued that to avoid “repetition and contradictions”, companies should be considered compatible with the safety rules at the state level as long as they meet federal or European standards.

“Several Bills of Amnesty International today – such as the proposals in California and New York – risked” through a line by violating the constitutional borders on how the states organized trade between the states, said the head of artificial intelligence policy, Andresson Horwatz and the head of legal officials.

The founders participating in the A16Z had previously indicated the organization of technology as one of the factors that led them to support Donald Trump’s attempt for a second period. The Trump administration and its allies later called for a 10 -year ban on the state’s artificial intelligence list.

Meanwhile, Antarbur came out in favor of SB 53.

“We have long said that we prefer a federal standard,” Jacques Clark, co -founder of the Arthur, said in a post. “But in the absence of this, it creates a solid plan for artificial intelligence that cannot be ignored.”