#United States #Salt Lake City #Glacis Technologies #OVERT #Utah AI Office

Glacis Technologies Partners with Utah for AI Accountability

On October 5, 2026, Glacis Technologies Inc. unveiled an agreement with Utah's Office of Artificial Intelligence Policy (OAIP) to make the use of independently verifiable AI receipts a necessity for vendors and systems participating in the state’s Artificial Intelligence Learning Laboratory (AILL). This landmark memorandum represents a crucial step in enhancing transparency and trust in the application of AI within health care.

The Importance of AI Receipts

Under the terms of this memorandum, every AI inference and associated evaluation of safety or governance must generate a cryptographically signed receipt. These receipts will serve as an evidence framework that enables independent evaluation and audit, reinforcing the accountability of AI systems in clinical settings. Joe Braidwood, founder and CEO of Glacis, emphasized, “Clinicians and patients need to be able to verify that AI safeguards are actually running. Utah's agreement provides a concrete evidence requirement, ensuring that a receipt is generated for every AI decision.”

Operational Implications

The oversight of this initiative falls under the purview of the OAIP, which will identify the specific vendors, systems, and events covered by this agreement. To implement this, AI workflows related to designated services will need to be equipped to generate and deliver the required receipts. This procedure plays a pivotal role as Utah expands its clinical AI program, striving to ensure that health care organizations, advisors, and independent evaluators collaborate effectively in assessing and validating various pilot projects.

For instance, if a workflow requires an AI to defer user requests to a human clinician, the receipt generated will document the AI’s operational control and decisions made during the event. This record—an independent and signed document—will provide evaluators with vital insights into how the AI system functions in real-world scenarios, further enhancing clinical oversight and evaluation of healthcare practices.

Collaborating for a Shared Future

Glacis is not alone in this initiative; the Coalition for Health AI (CHAI) has also joined the fray as a co-steward of OVERT, an open standard designed for the verifiable evidence of AI operations. “CHAI is excited to participate alongside our partner Glacis in this initiative,” remarked Brenton Hill, head of operations and general counsel at CHAI. The coalition strives to advance ethically responsible health AI practices by serving as a facilitator for independent evaluators involved with the sandbox program.

The verification process enabled by these receipts allows reviewers to confirm the integrity of records linked to a signing key. Assessing whether a referral reached the right clinical professional, if the AI’s control was effective, and whether the provided care was relevant are all integrated components in the evaluation process. However, it’s crucial to note that while the receipts provide evidence for clinical scrutiny, they do not guarantee safety, efficiency, or legal compliance on their own.

Keeping Data Secure

An important aspect of the agreement is that it facilitates the linking of evidence to covered events without necessitating the disclosure of sensitive data such as model weights, raw prompts, or underlying personal information unless otherwise agreed to in writing. This stipulation is pivotal for maintaining patient confidentiality while ensuring regulatory compliance.

As part of the initiative, the receipts will utilize hashes or other non-content identifiers to establish necessary connections while preserving privacy and data security.

Looking Ahead

The software standard, known as OVERT (Observable Verification Evidence for Runtime Trust), is an innovation developed by Glacis Technologies to ensure the verifiable evidence of AI operations. The specifications for OVERT are publicly available and governed under a royalty-free patent agreement. This collaboration with CHAI and the AIGovOps Foundation signals an inclination towards collective governance, assuring stakeholders of a high standard in AI deployments.

As the demands for transparency and reliability heighten in the AI landscape, the agreement between Glacis Technologies and Utah's OAIP sets a significant precedent. By establishing rigorous evidence protocols, stakeholders are promoted to embrace more reliable and accountable AI practices that benefit organizations, clinicians, and patients alike.

In conclusion, the transformative partnership between Glacis and Utah exemplifies how leadership in technology can foster trust and accountability, paving the way for broad implementation of AI in clinical settings—for the betterment of both healthcare providers and the patients they serve.