Oregon Gov. Tina Kotek issued an executive order on Sept. 23 directing the state to develop new procurement safeguards for advanced artificial intelligence systems, including standards for independent safety reviews and an assessment of whether the state should require a “kill switch” for certain AI models.
The executive order establishes a state policy of seeking to purchase, contract for, or otherwise use frontier AI models that have received independent, third-party safety reviews, consistent with state and federal law. The order took effect immediately.
Under the order, Oregon’s state chief information officer has 90 days to develop and submit an implementation proposal to Kotek. The governor will reassess the need for the executive order every three months, and the order will remain in effect until she terminates it.
“While oversight and regulation would be strongest and most effective at the federal level, President Trump continues to dismiss the need for even the most basic guardrails,” Kotek said in a press release. “States, in the meantime, must act with more urgency and do everything possible to manage how AI is impacting our communities.”
The implementation proposal must define a “frontier artificial intelligence model” for state procurement and use and establish standards for determining whether a model has received an adequate independent safety review. It must also develop methods for assessing the independence, qualifications, scope, methodology, and transparency of third-party reviewers.
The order further directs the state CIO to evaluate the viability of making a “kill switch” one of Oregon’s procurement standards for frontier AI models. The order does not itself impose such a requirement.
Other elements of the proposal must address how agencies would incorporate AI safety reviews into procurement, contracting, and technology-use practices, as well as possible exceptions when third-party reviews are unavailable. Those exceptions could cover emerging technologies, specialized applications, or state or public safety needs.
The CIO must also recommend procedures for ongoing monitoring and reevaluation as AI models, risks, and independent review standards evolve. The proposal will address transparency and accountability for state procurement and use of frontier AI and identify any statutory, administrative, budgetary, or policy changes needed for implementation.
Once Kotek approves and directs implementation of the proposal, the CIO will coordinate with state agencies to put the policy into effect.
The executive order comes as leading AI researchers and industry executives have increasingly warned about the technology’s potential risks, while other states have begun establishing their own guardrails for AI development and use.
“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.
“We must act with urgency to address this issue, and I am committed to doing everything in my executive authority to put safeguards in place until the Oregon Legislature and Congress can pass comprehensive standards for AI companies that wish to bring their products to market here,” she added.
