Oregon sets review requirements for frontier AI procurement

Oregon Gov. Tina Kotek issued Executive Order No. 26-26 to establish safeguards for how state government buys, contracts for and uses frontier AI models. The order says Oregon will use models that have received independent, third-party safety review, consistent with state and federal law, while continuing to support innovation and responsible AI use.

State CIO Terrence Woods must develop standards or criteria for adequate third-party AI safety review and return an implementation proposal to the governor in 90 days. Oregon officials said the framework must define frontier AI for state procurement and use, assess whether a “kill switch” requirement is viable, and integrate safety reviews into procurement, contracting and technology-use practices.

Oregon is looking to other states as it develops its approach. California recently enacted Senate Bill 813 for independent assessments of AI safety risks and Assembly Bill 1405 to create a registry for AI auditors, while New York plans to require AI developers to register with the state in November and meet transparency standards in January 2027, including reporting critical safety issues within 72 hours.