The Department of Government Efficiency (DOGE) team working at the Department of Housing and Urban Development (HUD) used artificial intelligence to help shape housing policy, but the agency is now blocking Freedom of Information Act (FOIA) requests for documents detailing how the AI tools were developed and used, according to documents obtained by Democracy Forward, a nonprofit legal organization.
The DOGE Team at HUD
Last year, WIRED reported that Christopher Sweet, then a third-year economics student at the University of Chicago, had joined the DOGE team at HUD alongside Scott Langmack, who came from the property technology startup Kukun. According to HUD employees who spoke to WIRED, Sweet’s primary focus was using AI to identify agency rules for potential rescission or contract cancellations, part of a broader government-wide effort. HUD staffers told WIRED that employees were asked to give feedback on regulations flagged by the AI for removal, though some described the effort as redundant.
Sweet graduated in June with a degree in economics; Langmack is now executive director of deregulation AI at the Office of Management and Budget (OMB), according to his LinkedIn. Neither Sweet, Langmack, HUD, OMB, nor the White House responded to requests for comment.
FOIA Denials and Withheld Documents
Democracy Forward requested more than 100 documents about HUD’s use of AI for decision-making. HUD refused to release them, citing a “nonexistent AI privilege” and a real but narrow “presidential communications privilege” that generally applies only to the president and their immediate advisers. Several withheld document titles, listed in the FOIA response, suggest the DOGE team was using AI to shape policy. One document, labeled “GPT defined Econ Analysis approach 11 10 25.docx” and belonging to Langmack, was exempted as “deliberative AI input.” Another, “RegulatoryAnalysisPrompt.pdf,” also Langmack’s, indicates the team was crafting prompts for regulatory analysis. Other withheld documents were labeled as “regulatory analysis” for various HUD programs, though it is unclear if AI was involved in their creation.
Expert Concerns
Tori Noble, a staff attorney at the Electronic Frontier Foundation, called the lack of transparency particularly worrisome because AI tools can hallucinate, show bias, or simply get things wrong. “It’s not necessarily the case that we'd always know how tools are being used,” she said. “So having access to the prompts is really the best way to be able to tell what officials are using these tools for and how harmful those uses might be.”
Currently, no US law requires the government to disclose if AI was used in creating rules, policies, or regulations. Mark Fagan, a lecturer at the Harvard Kennedy School, noted that if AI is used to assess policy, “at this stage in the development and use of AI, it is good protocol to indicate that” to build confidence. However, he added that AI used for internal deliberation—like Googling how others handled a policy—might not warrant disclosure.
Implications for Enterprise Buyers
For enterprise technology leaders, this case underscores critical risks when AI informs high-stakes decisions without oversight. The same lack of transparency—no mandated disclosure, no requirement to share prompts or training data—could surface in procurement of AI tools for supply chain, logistics, or trade compliance. Without clear audit trails, enterprises risk deploying AI that produces biased or erroneous outputs, with legal and reputational consequences. The absence of laws like those in the EU’s AI Act means buyers must demand their own transparency standards from vendors—something DOGE’s use at HUD illustrates is both possible and dangerous when absent.
