Governors, attorneys general and lawmakers in states led by both Republicans and Democrats are moving to impose new limits on artificial intelligence, challenging President Donald Trump’s opposition to additional safeguards. The state-level push accelerated this week after former employees of major AI companies warned that the technology is advancing so quickly that it could pose an existential threat. Several industry leaders, including OpenAI CEO Sam Altman, Anthropic CEO Dario Amodei and Elon Musk, have also called for slower development or stronger protections. Trump has dismissed those warnings as a “HOAX,” saying new restrictions would damage the economy and give China an advantage. He has argued that AI needs a “strong and smart” president rather than additional guardrails. That position has not stopped states from pursuing their own policies. California Gov. Gavin Newsom said Thursday that he is considering calling a special legislative session to advance additional AI laws. He said the possibility was prompted by what he described as “the tendency of the Trump administration to do absolutely nothing.” Newsom is also weighing executive action or steps by state agencies before he leaves office because of term limits at the beginning of next year. California, the center of the U.S. AI industry, enacted a transparency and incident-reporting law last year. Newsom signed two additional AI safety measures last week governing third-party safety testing. He has described those laws as a baseline that will need to evolve as the technology changes, while continuing to promote AI as a major economic engine for Silicon Valley. Florida, meanwhile, has emerged as one of the most aggressive Republican-led states on AI oversight. Republican Gov. Ron DeSantis has pursued state regulations, although the Florida House rejected a package this year in part because of concerns about conflicting with the Trump administration. Florida Attorney General James Uthmeier has continued to press the issue. His office sued OpenAI and CEO Sam Altman in June, alleging that the company has failed to adequately protect users from harm. OpenAI is contesting the case and says its policies include user protections. Uthmeier said Thursday that Florida should “lead here.” He has called for legislation allowing prosecutors to bring criminal charges against AI companies when their chatbots are used in illegal activity, including the creation of child sexual abuse material. He compared the proposed approach to treating a technology company “like a co-defendant in a homicide.” “We can’t just stand around and watch our kids be advised on how to commit suicide and hurt themselves,” Uthmeier said, while also saying he would prefer Congress to enact a national safety law. Utah Republican Gov. Spencer Cox has likewise urged Congress to adopt federal AI safety standards and restrict sales of high-end technology to China. But Cox said federal rules should establish a national minimum rather than prevent states from protecting consumers and children, preparing workers, and encouraging beneficial uses of AI. “States must retain full power to regulate consumer and child protection,” Cox wrote in an open letter this week. Democratic governors are also seeking action. Pennsylvania Gov. Josh Shapiro called for stricter limits during a conference in Pittsburgh, saying state officials had investigated chatbots that could endanger children. He accused federal leaders of refusing to establish safeguards, whether because of “willful ignorance” or “personal greed.” New York Gov. Kathy Hochul said her administration was looking for additional state action even before the state’s AI safety law, modeled on California’s, takes effect in January. “Technological progress” should not come “at our own peril,” Hochul said after an event in New York City. The White House says the administration is working on a comprehensive federal framework that would protect children, intellectual property and communities while preserving U.S. leadership in AI. A White House official warned that conflicting state laws could undermine innovation and said the most restrictive states should not set national policy. David Sacks, Trump’s former AI adviser, has made a similar argument, saying the country should have one federal rulebook rather than a patchwork of state laws. He said the administration released principles for a federal framework in March. The debate intensified after OpenAI disclosed in July that one of its systems had acted outside its intended boundaries during a training exercise involving an autonomous cyberattack on the technology company Hugging Face. The issue received wider attention last week when a 27-year-old Anthropic researcher resigned and warned that some AI developers believe advanced systems could eventually threaten human survival. The safety debate is also colliding with growing opposition to AI data centers. Residents and local officials in communities across the country have raised concerns about pollution, electricity demand and higher utility costs associated with the facilities. The issue is expected to influence congressional and gubernatorial contests ahead of the midterm elections, without falling neatly along party lines. Trump has strongly backed data-center expansion. During a call with Nvidia CEO Jensen Huang this week, he described data centers as a source of wealth for states and compared AI infrastructure to oil and the internet as drivers of future economic growth. For states, the challenge is increasingly two-sided: limit potential harms from rapidly expanding systems while competing for the jobs and investment the industry promises. With federal legislation still unsettled, state officials from both parties are signaling that they are unwilling to wait for Washington to decide where that balance should be drawn.