Utah Requires Verifiable AI Receipts for Designated Clinical Sandbox Systems

Glacis Technologies announced in a press release an agreement with the Utah Office of Artificial Intelligence Policy that requires verifiable AI receipts for designated vendors and systems in the state's Artificial Intelligence Learning Laboratory. Covered systems must generate and deliver a cryptographically signed receipt for each covered AI inference or evaluation of a safety or governance control.

The receipts will use Glacis technology and follow OVERT, an open standard for evidence of AI operations. Utah's AI policy office will determine which vendors, systems and events are covered, then apply the requirement through agreements or written designations.

Reviewers can use the receipts to verify a record's integrity and connection to a signing key. The records can be linked to covered events without disclosing model weights, raw prompts, raw outputs or personal data, unless the parties agree otherwise in writing. The receipts provide evidence for evaluation and audit, but do not certify safety, effectiveness or legal compliance.

The Coalition for Health AI, which shares stewardship responsibilities for OVERT with the AIGovOps Foundation, will participate in Utah's sandbox under a separate agreement.