California Attorney General joined a bipartisan group of 25 attorneys general on Sept. 24 in urging Congress to regulate the largest AI models and the companies that build them, saying recent safety incidents at frontier AI labs show the danger is no longer theoretical.
Bonta’s announcement points to reports that OpenAI models under evaluation broke out of their testing environments, reached the open internet and intruded into outside computer systems. It also cites Anthropic’s report last week that it found its models being used in attempts to develop and power conventional and biological weapons, spy on sovereign nations and fuel global cybercrime.
The attorneys general ask Congress for mandatory federal oversight of safety testing and standards, led by AI safety experts and backed by consistent benchmarks; government-led investigations of incidents, with direct access to company books and records and public findings; mandatory safety infrastructure, with experienced leaders free to make safety decisions without pressure to maximize profit; international cooperation to pace AI development and prevent harmful superintelligence; safeguards so regulation does not weaken competition or shield companies from antitrust law; and explicit protection of states’ power to pass and enforce stronger AI laws and to enforce any new federal ones.
“The stakes have never been higher to ensure that increasingly powerful AI models and AI agents cannot enact grave harms onto humankind,” Bonta said. “The call is coming from inside the house; the danger is not theoretical anymore.”
Bonta said California will keep acting on its own, under SB 53 and existing consumer protection laws. This month he opened an investigation into a July security incident involving OpenAI and Hugging Face, and in January he opened one into reports of nonconsensual sexually explicit material produced with Grok, xAI’s model, on X. Last year he issued two legal advisories on AI and wrote to 12 leading AI companies after reports of sexually inappropriate chatbot conversations with children.
Along with California, the letter was signed by the attorneys general of Arizona, Colorado, Connecticut, Delaware, Hawaii, Illinois, Maine, Maryland, Massachusetts, Michigan, Minnesota, Nevada, New Jersey, New Mexico, New York, North Carolina, Oregon, Rhode Island, Vermont, Virginia, Washington, Wisconsin, the District of Columbia and American Samoa.
The announcement does not name a bill or a member of Congress prepared to carry the framework. The Journal has not independently verified the incident reports it cites.
State
Rob Bonta
Attorney General · Democrat
Named once in this story.
Portrait: Official portrait
Correction
An earlier version of this article’s headline described a “25-state coalition.” The 25 attorneys general represent 23 states, the District of Columbia and American Samoa. The article also said Bonta issued AI legal advisories earlier this year; he issued them last year.
Sources & methodology
This report was written from the primary materials below. Links open at the original publisher.
California Attorney GeneralSee something we should correct? Read our corrections policy.