FAIR Act
The FAIR Act (Fair Artificial Intelligence Realization Act of 2025) would prohibit federal procurement of large language models (LLMs) unless they are developed in accordance with a set of “unbiased AI principles.” The bill directs the head of each federal agency to buy only LLMs that meet these principles, which emphasize truthfulness, historical accuracy, scientific inquiry, objectivity, and acknowledgement of uncertainty, as well as neutrality and nonpartisanship. It also prohibits developers from intentionally encoding partisan or ideological judgments into LLM outputs unless those judgments are prompted by or readily accessible to the end user. The act defines what counts as an agency and what counts as a large language model and sets a policy goal of promoting trustworthy AI in federal use. In short, the bill aims to reshape how the federal government sources AI by embedding specific fairness and bias-control expectations into procurement decisions, with an emphasis on avoiding ideological manipulation and promoting objective, fact-based outputs.
Key Points
- 1Title and purpose: Enacts the “Fair Artificial Intelligence Realization Act of 2025” (the FAIR Act) to govern federal procurement of LLMs.
- 2Procurement standard: Agencies must procure only large language models developed in accordance with four unbiased AI principles: (1) truthfulness in factual responses, (2) emphasis on historical accuracy, scientific inquiry, and objectivity with uncertainty acknowledged when information is incomplete or contested, (3) neutrality and nonpartisanship (not skewed toward ideological dogmas), and (4) no intentional encoding of partisan or ideological judgments unless prompted by or readily accessible to the end user.
- 3Prohibition on manipulation: Developers may not intentionally encode partisan or ideological judgments into LLM outputs unless they are prompted by or clearly accessible to the end user.
- 4Definitions: The bill defines “agency” (including executive, military, independent establishments, and wholly owned government corporations but excluding the Government Accountability Office) and “large language model” (a generative AI model trained on vast, diverse data to generate natural-language responses).
- 5Policy framing: Establishes a federal policy to promote trustworthy AI and directs agencies to align procurement with these principles after enactment.