Skip to content
Center for AI Oversight
The public record

Comments & Submissions

The Center’s recommendations submitted through formal channels to standards bodies and regulatory organizations.

Status as of September 30, 2026: both proposals have been submitted. Neither has been accepted or adopted by the receiving organization. These documents state the Center's positions.
National Association of Insurance CommissionersSubmitted

Exhibit B Is the Oversight Exhibit

Comment on the AI Risk Evaluation Supplement, Version 5.0

Document date
September 26, 2026
Submitted to
Big Data and Artificial Intelligence (H) Working Group, NAIC
Version addressed
Version 5.0, exposed August 31, 2026
Author
Brian J. Allen, Esq., Executive Director

The comment identifies Exhibit B as the Supplement's oversight exhibit and proposes additions that make the object of oversight, the accountable parties, and the decisions on record more explicit.

Five recommendations address program scope; accountability, exclusions, and competency; ratified risk tolerance; authority to operate; and escalation and disclosure. The companion document supplies proposed checklist language, with one amendment, seven new entries, and one optional assurance item.

The full comment also addresses provisions in Exhibit C and the senior officer attestation in the separate third-party vendor instrument. The proposed additions are the Center's text, not amendments adopted by the NAIC.

National Institute of Standards and TechnologySubmitted

The Missing Oversight Layer in the Trustworthy AI in Critical Infrastructure Profile

Comment to the TACIP Community of Interest

Document date
August 11, 2026
Submitted to
Raymond Sheh and Martin Stanley, NIST
Version addressed
Community of Interest Discussion Draft, August 3, 2026
Author
Brian J. Allen, Esq., Executive Director

The Center proposes a chartered oversight program that identifies what an organization governs, who is accountable, who authorizes a deployment, and who can withdraw that authority.

The comment supplies proposed Practice 13 text in the format of the August draft, indexed comments, responses to reviewer questions, and a comparison of accountability mechanisms across critical infrastructure sectors. A separate two-page summary presents the central finding and proposal.

Version context. This submission addresses the August 3 discussion draft. The September 13 draft uses Practice 13 for retiring AI systems and withdrawing their authority. That numbering does not establish acceptance of the Center's proposed practice. The original submission is preserved here.

Both submissions apply the Center's position on the oversight layer. For a briefing on this work, contact the Center.

The AI Oversight Brief

A short briefing on a consequential oversight decision.