Utah’s Cox Directs State Agencies to Use AI With Human Oversight

Utah Gov. Spencer J. Cox has signed an executive order directing state agencies to pursue AI-driven improvements while keeping people involved in consequential decisions. The governor’s announcement calls for employee training, security safeguards and transparency, including human review of automated determinations affecting individual rights. It puts implementation duties on agencies, not just principles on paper.

Better tools, retained authority

Executive Order 2026-08 advances Utah’s Pro-Human AI Initiative, established in December 2025. It applies to executive-branch agencies, including the State Tax Commission, National Guard and Board of Pardons and Parole. The initiative is the existing policy framework; this order supplies directions for using AI within state government.

The order distinguishes two goals. “Human-guided” systems preserve human agency and authority while protecting dignity. “Human-enhancing” systems give workers better tools for solving problems and strengthening communities. The intended relationship is assistance under human control, with workforce development built into the approach.

Agencies must evaluate cost-effective uses for internal operations and investigate automated services for residents. For determinations affecting individual rights, the governor’s release specifies human-in-the-loop review: a person remains involved in the decision. Agencies must also disclose AI inputs and its role in decision-making.

The choice between innovation and safety in government is a false one.

Utah Gov. Spencer J. Cox, in the governor’s news release

Training and defenses get named owners

The Department of Government Operations will lead AI-driven process improvements intended to improve efficiency and decision-making. It must also establish technical safeguards and guidance against cybersecurity threats and rogue AI agents. That assigns the department both an adoption role and a responsibility to protect government systems.

The Division of Technology Services and Division of Human Resource Management must develop and implement an AI-literacy training strategy for state employees. Agency AI leads will receive specialized training. Technology Services must also coordinate uniform implementation standards and evaluate AI tools for detecting and responding to cybersecurity threats.

Digital identity is another part of the security work. Government Operations is directed to implement the State-Endorsed Digital Identity Framework where applicable. The governor’s release describes it as infrastructure for cryptographically verifiable digital credentials—credentials whose authenticity can be checked using cryptography—and privacy protection.

A reporting deadline, not a results claim

Agencies are directed to collaborate and share information as they evaluate AI uses. Government Operations must submit an annual progress report to Cox, with the first due July 1, 2027. That is a future accounting requirement, rather than evidence that the order’s intended improvements have already been delivered.

Government Technology reports that the order took effect with Cox’s signature. The exact signing date needs clarification: the governor’s release carries an October 6, 2026 header but an October 30 dateline, while Government Technology’s October 7 article says it was signed Tuesday.