Oregon Governor Issues Executive Order for AI Safety Reviews in State Government
Summary
- • Oregon Gov. Kotek orders AI safety review process for all state government AI use
- • Third-party reviewers required before state deploys certain AI technologies
- • CIO has 90 days to define 'frontier' AI models and propose regulations
- • Follows California's similar EO; cites Trump administration's federal inaction
Details
Executive Order Signed by Gov. Kotek
Gov. Tina Kotek directed Oregon government to develop a comprehensive AI regulation framework for state use, citing growing safety risks and federal inaction under the Trump administration.
Third-Party AI Safety Review Required
State must develop qualifications for external reviewers to assess safety of certain AI technologies before state deployment; specific technologies subject to review remain to be defined.
Frontier AI Models to Be Defined
Oregon's CIO is ordered to create a formal definition for 'frontier' AI models — the most advanced systems — which will determine which technologies are subject to the stricter regulatory requirements.
Kill Switch Feasibility Study Ordered
The executive order requires an assessment of whether a 'kill switch' for frontier AI systems is technically viable. Technology experts note this would present immense implementation challenges.
90-Day Regulatory Deadline for CIO
Oregon's chief information officer has 90 days from the order's signing to develop and submit proposed regulations, meaning concrete definitions could arrive by late December 2026.
California Issued Similar EO One Week Prior
California Gov. Gavin Newsom signed a nearly identical executive order the prior week, suggesting a coordinated or parallel state-level approach to AI governance is forming across blue states.
Federal Regulatory Vacuum Cited as Driver
Kotek explicitly cited President Trump's resistance to federal AI regulation as the motivation: 'States, in the meantime, must act with more urgency and do everything possible to manage how AI is impacting our communities.'
Policy actions and contextual developments from Oregon's AI executive order
What This Means
Oregon's executive order adds another state to a growing coalition taking unilateral action to govern AI in the absence of federal policy. The order's three-pronged approach — mandatory safety reviews, definitional standards for frontier AI, and kill-switch research — closely mirrors California's recent effort, suggesting an emerging de facto state-level governance template. If similar orders multiply, they could create compliance fragmentation for AI vendors deploying tools to government clients across different states. With a 90-day deadline, concrete regulatory definitions for frontier AI could emerge from Oregon by late December 2026.