Oregon Is Latest State to Move on Frontier AI Regulation

States including California, New York and now Oregon are moving to regulate frontier AI models, taking notable steps given the federal government’s hesitancy to lead in regulatory oversight.Oregon Gov. Tina Kotek on Wednesday issued Executive Order No. 26-26, “Establishing Responsible Artificial Intelligence Procurement Standards for State Government,“ to set safeguards and oversight around AI tools.While supporting innovation and responsible AI use, the state will buy, contract for and use “frontier artificial intelligence models that have received independent, third-party review for artificial intelligence safety” consistent with state and federal law, according to the EO.

The EO directed state CIO Terrence Woods to “develop standards or criteria for adequate third-party review for AI safety,” according to a news release. Woods will come forward in 90 days with an “implementation proposal” for the governor, said Hope Hiebert, communications director at Oregon Enterprise Information Services.“That proposal will address, among other questions, the standards or criteria for adequate independent, third-party safety review,” Hiebert said in an email, noting “we do not have a separate timeline to share yet for adopting any resulting standards.”Oregon will likely look to other states like California for guidance as the state develops an AI regulatory framework, officials said.

Per the EO, it must address issues including a definition of frontier AI appropriate for state government procurement and use; evaluate the viability of a “kill switch” requirement as a standard for frontier AI models; and set processes for including AI safety reviews in “state procurement, contracting and technology-use practices.”In the news release, the state took note of differing postures on AI and its regulation.“As AI industry leaders call for a slowdown on AI development, President [Donald] Trump and congressional leaders continue to deny the need for independent oversight or even basic regulatory safeguards as frontier AI models continue to demonstrate the potential to break containment and threaten cybersecurity infrastructure,” the state said.The work set out in the EO is similar to steps being taken by California to put in place an independent, third-party review and evaluation structure for frontier AI models and systems which are becoming increasingly used by both the private and public sectors in the state. (“Frontier” refers to the most cutting-edge AI tools and products.)Earlier this month California Gov. Gavin Newsom signed state Senate Bill 813, establishing a framework for independent, third-party assessments of potential safety risks posed by AI systems, and Assembly Bill 1405, companion legislation creating a state registry for AI auditors and establishing standards for their independence and transparency.On Sept. 18, Newsom signed an executive order to fast-track the legislation which raised the idea of a “kill switch” for frontier AI models.“Artificial intelligence has the potential to improve our lives, from boosting productivity to making scientific and medical breakthroughs, but without effective guardrails, AI also presents substantial and possibly catastrophic risks,” said California state Sen. Jerry McNerney, D-Pleasanton, in an email. McNerney was the author of SB 813.“SB 813 is a strong first step, but it’s clear there is more we need to do,” he said. “AI safety standards are an idea I have been pushing for years. Standards are commonplace across industries. There are building and construction standards, airline standards, electrical standards, and banking standards, to name a few."These steps taken by states come as even AI technology leaders sound the alarm about the possibly dangerous reaches of the technology."It’s encouraging to hear AI companies are beginning to come forward with proposals to address safety concerns, but we cannot simply wait for the private sector to self-regulate a technology that poses such an existential and evolving risk," Kotek said in a statement.Meanwhile, New York is also developing AI regulations to require AI developers to register with the state in November. In January 2027, AI companies must “comply with strict transparency standards,” said New York Gov. Kathy Hochul in a Sept. 21 press briefing. This includes publishing safety protocols and reporting critical safety issues within 72 hours. The move is part of the state’s AI safety law known as the Responsible AI Safety and Education Act, passed last year.“I know there’s responsible AI developers out there who care deeply about safety transparency and doing things the right way. Developers who want common-sense regulation, who want uniformity, who want to work with us in good faith,” Hochul said, stressing the need for a unified AI regulation strategy. “It can’t just be New York. It can’t just be California. It can’t just be a handful of governors. We need a national and global strategy.”

Skip Descant writes about smart cities, the Internet of Things, transportation and other areas. He spent more than 12 years reporting for daily newspapers in Mississippi, Arkansas, Louisiana and California. He lives in downtown Yreka, Calif.