California Governor Faces September 30 Deadline to Decide on Bill Imposing Oversight on Major AI Models

California Governor Gavin Newsom has until September 30 to make a decision on bill SB 1047, which was passed by the California legislature and obligates companies developing frontier artificial intelligence models to implement mandatory safety protocols and submit their models to independent auditing.

The law targets models that exceed specific thresholds in terms of compute size and training costs, requiring companies to conduct risk assessments prior to deployment and take measures to mitigate these risks, while granting authorities the right to investigate potential harms arising from these models.

Stances on this legislation vary sharply; ethical advocacy groups and some academics seeking to cap risks support it, while many Silicon Valley leaders oppose it, viewing it as an obstacle to innovation and a threat to California's tech leadership.

If the governor signs the law, California will become the first U.S. state to impose direct regulatory requirements on the development of frontier AI models, setting a precedent that could inspire similar legislation at the federal level or in other states.

What do these terms mean?

Legislative Veto (Veto): The governor's right to reject a law passed by the legislature and prevent its implementation, returning the bill to the legislature, which would then require a special majority to override the rejection.