Governor Tina Kotek issued an executive order September 23 directing Oregon to develop additional safety standards for advanced artificial intelligence systems purchased or used by state agencies. The statewide order applies to executive-branch agencies serving residents across Oregon, including those in Southern Oregon.
Executive Order 26-26 directs the state’s chief information officer to develop standards for the safety review of frontier AI models, a term used for highly capable AI systems. The review would be conducted by an independent third party rather than relying solely on a vendor’s assessment. The order also calls for the state to examine whether it would be practical to require a way to shut down such a system.
Neither requirement has been fully put into operation. The chief information officer has 90 days to submit an implementation proposal to the governor. Oregon will then have to determine how the proposed standards would work in state purchasing and agency use. The order took effect immediately, remains in place until the governor ends it and must be reconsidered every three months.
For people in Josephine, Jackson and other Southern Oregon counties, there is no new application step, identification requirement or deadline because of this order. It does not change who qualifies for a state benefit or announce that AI will make decisions about individual residents. Its immediate effect is on state government’s work evaluating technology that agencies may use.
That work could eventually matter wherever Oregonians interact with a state agency, whether at a local office or through a statewide website. An agency considering an advanced AI tool would have to account for the state’s procurement and safety standards as they are developed. The order does not identify a particular Southern Oregon office, service or AI system that will change.
Oregon had begun setting rules for agency AI use before this week. Kotek created a State Government Artificial Intelligence Advisory Council in 2023, and the council delivered its recommended action plan in February 2025. The state says executive-branch agencies already operate under requirements addressing risk, human accountability, transparency and oversight. Current state policy also requires agencies to review proposed AI uses through an information technology approval process, including AI features added to software they already have.
The new order focuses on the most advanced models and how their safety should be assessed when the state considers using them. It also puts the proposed shutdown capability under review. The governor has directed officials to study its feasibility; she has not declared that every state AI system must have one.
For now, the next step is a proposal from state officials on how the added safeguards will work. Whether residents notice a change in any service will depend on that proposal and on future agency decisions about the technology they purchase and use.