Government Withholds Records on AI Use in Housing Policy, Raising Transparency Concerns

Summary

Members of the Department of Government Efficiency (DOGE) deployed at the Department of Housing and Urban Development (HUD) used artificial intelligence tools, including GPT-based systems, to inform policy decisions on regulatory rescission and program analysis. According to documents obtained through Freedom of Information Act requests by the nonprofit legal organization Democracy Forward, the agency has systematically denied disclosure of how these AI systems were developed and used to shape housing policy decisions.

HUD has withheld over 100 documents related to AI use, citing questionable legal justifications including a non-existent "AI privilege" and presidential communications privilege—which typically protects only direct communications with the president and immediate advisers. Documents with names like "GPT defined Econ Analysis approach 11 10 25.docx" and "RegulatoryAnalysisPrompt.pdf" indicate the DOGE team was using AI tools to conduct regulatory analysis and identify potential policy changes.

Experts and watchdog groups warn that lack of transparency around government use of AI in policymaking poses significant risks. Because AI systems are known to hallucinate, exhibit bias, and produce inaccurate information, understanding how they influence government decisions is critical to maintaining policy integrity and public trust. Notably, the U.S. has no laws currently requiring government agencies to disclose whether AI was used in creating or modifying regulations and policies.

• AI transparency advocates argue that disclosure of AI use in government is essential, given documented risks including hallucinations, bias, and factual errors