Government Used AI for Housing Policy, Won’t Reveal Details

The federal government deployed artificial intelligence to shape housing policy, but now refuses to disclose how those tools were developed or what decisions they influenced. The Department of Housing and Urban Development (HUD) is blocking Freedom of Information Act (FOIA) requests for records related to the AI systems used by members of the so-called Department of Government Efficiency (DOGE), according to documents obtained by Democracy Forward, a nonprofit legal organization.

Last year, WIRED reported that Christopher Sweet, then a third-year economics student at the University of Chicago, joined the DOGE team at HUD alongside Scott Langmack, who came from the property technology startup Kukun. Sweet’s primary role, according to HUD employees who spoke to WIRED at the time, involved using artificial intelligence to identify agency rules for potential rescission or contract cancellations, part of a broader government-wide effort. Some HUD staffers said they were brought in to provide feedback on regulations flagged by the AI, while others described the effort as redundant.

Sweet graduated from the University of Chicago in June with a degree in economics. Langmack now serves as the executive director of deregulation AI at the Office of Management and Budget (OMB) under the Executive Office of the President, according to his LinkedIn profile.

Democracy Forward requested more than 100 documents related to HUD’s use of AI for decision-making, but the agency withheld them all. Among the reasons HUD cited for not releasing the records were a nonexistent AI privilege and a presidential communications privilege that is real but generally applies only to the president and their immediate advisers. Several withheld documents, whose names appear in the FOIA but whose contents remain unknown, suggest that the DOGE team at HUD used AI tools to help make policy decisions. Sweet, Langmack, HUD, OMB, and the White House did not respond to requests for comment.

One document, labeled “GPT defined Econ Analysis approach 11 10 25.docx” and belonging to Langmack, was exempted because it was designated as “deliberative AI input.” Another document, titled “RegulatoryAnalysisPrompt.pdf,” also Langmack’s, indicates the DOGE team was exploring how to create prompts for conducting regulatory analysis. Several other withheld documents, labeled as some form of “regulatory analysis” for different HUD programs, were exempted as part of the deliberative process, though it is unclear if AI was used in their creation.

Tori Noble, a staff attorney at the Electronic Frontier Foundation, warns that the lack of transparency around how AI tools might be used to create or change policy is especially troubling, given that such tools are known to 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 says. “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 U. S. law requires the government to disclose whether AI was used in the creation of rules, policies, or regulations. Mark Fagan, a lecturer at the Harvard Kennedy School, says that if AI is being used to assess policy as one tool in the toolkit, it would be good protocol to indicate that, in part to build confidence in the use of AI in government.