FAIR Act

The Fair Artificial Intelligence Realization Act of 2025 (FAIR Act) establishes a policy for the United States to promote the innovation and use of trustworthy artificial intelligence . To achieve this, the bill prohibits federal agencies from procuring large language models unless they are developed in accordance with specific unbiased AI principles. These principles require large language models to be truthful in response to factual prompts, prioritize historical accuracy , scientific inquiry, and objectivity, and acknowledge uncertainty when information is incomplete. Furthermore, the bill mandates that large language models must function as neutral, nonpartisan tools , explicitly stating they should not manipulate responses in favor of ideological dogmas such as diversity, equity, and inclusion. Developers of these models are also prohibited from intentionally encoding partisan or ideological judgments into the output, unless such judgments are specifically prompted by or readily accessible to the end user. The Act defines "agency" broadly to include most federal departments and establishments, and "large language model" as a generative AI model trained on vast datasets to produce natural-language responses.