New OMB Memo Imposes Detailed New Disclosure Requirement on Government Contractors Utilizing LLMs

A new memo from the Office of the Management of the Budget imposes detailed disclosure requirements on government contractors using large language models (LLMs), a development published on 23 December 2025.

According to the Office of the Management of the Budget, when procuring LLMs, government agencies must require vendors to provide information so they can determine compliance with Unbiased AI Principles and that the product is not in violation of Executive Order 13960, “Promoting the Use of Trustworthy Artificial Intelligence in the Federal Government”. The memo aims to prevent LLMs from running afoul of two Unbiased AI Principles—Truth-seeking and Ideological Neutrality—by requiring minimum disclosures.

At a minimum, agencies are requires to request from vendors an Acceptable Use Policy, Model, System, and/or Data Cards, End User Resources, and a Mechanism For End User Feedback. These measures are intended to help assess training processes, risks and mitigations, and outputs that violate the Unbiased AI Principles when government procurement occurs.