TLDR:
California’s AI legislation, introduced by State Senator Scott Wiener, has cleared the Senate and Assembly Appropriations Committee. The bill mandates safety testing for advanced AI models and empowers the California Attorney General to file charges against companies for causing harm. The legislation has faced opposition from tech companies and lawmakers, with concerns about stifling innovation. Industry-friendly amendments have been made, but the bill’s fate remains uncertain pending a vote in the full California State Assembly and Governor Newsom’s decision.
Key Points:
- Bill to regulate AI companies in California clears Assembly Appropriations Committee
- Mandates safety testing for advanced AI models and empowers Attorney General to file charges
A bill to regulate artificial intelligence companies in California has cleared a crucial hurdle in the state’s Assembly Appropriations Committee but remains far from being a sure thing to pass. The measure cleared the California State Senate earlier this year in a 32-1 vote, and now the legislation will head for a vote before the entire state assembly. The date for when that vote could occur has yet to be set, but the state legislative session ends on August 31. So, there is a small window left for the bill to clear the legislature. Then, if passed out of the legislature, it will head to California Governor Gavin Newsom (D) who will have until the end of September to decide to sign or veto the bill.
The proposal, from California State Senator Scott Wiener (D), would mandate safety testing for the most advanced AI models, a threshold defined by those that cost more than $100 million to develop or meet a certain threshold of computing power. These safety tests would be required before the models could be released publicly. In addition, the California Attorney General would be empowered to file charges against companies if their technologies were found to have caused serious harm.
The legislation has garnered pushback from many tech companies, including some of the largest tech giants like Google, Meta, Anthropic, and OpenAI, and venture capital funds such as a16z and Y Combinator. Critics argue that the legislation could stifle innovation or push AI development out of California if the regulations are too onerous, mainly due to the amount of liability companies would have to deal with under the requirements.
In the state’s Assembly Appropriations Committee’s markup, the legislation received industry-friendly amendments that may impact its chances of passing. The most important tweak appears to be a language adjustment to “reasonable care” from “reasonable assurance.” Wiener has accepted the amendments, but it is uncertain whether these revisions will go far enough to appease the opposition from AI companies, investors, and lawmakers.