DOGE’s work inside the Department of Housing and Urban Development has moved beyond access to agency systems and into a more consequential area: rewriting rules. According to the source article, Christopher Sweet has been tasked with using artificial intelligence to review HUD regulations, compare them with underlying laws, and identify places where rules could be loosened or removed.
Who is leading the AI review
Christopher Sweet was introduced to HUD staff earlier this month as originally from San Francisco and most recently a third-year at the University of Chicago, where he was studying economics and data science. The University of Chicago confirmed to WIRED that Sweet is “on leave from the undergraduate college.”
In an email reviewed by WIRED, DOGE staffer Scott Langmack described Sweet as joining the HUD DOGE team with the title of special assistant. Langmack wrote that “a better title might be ‘Al computer programming quant analyst,’” and added that Sweet speaks Portuguese fluently because of family roots from Brazil.
Sweet did not respond to requests for comment about his work. HUD, asked to clarify his role, said through a spokesperson that it does not comment on individual personnel.
What the AI project is doing
Sweet’s central assignment appears to be an AI-driven review of HUD rules and regulations. The project is described as comparing agency regulations with the laws that authorize them, then identifying areas where the agency may have gone beyond what is required.
One focus is regulation connected to the Office of Public and Indian Housing, known as PIH. Sources told WIRED that Sweet created an Excel spreadsheet with around a thousand rows of policy areas where the AI tool flagged possible agency “overreach” and suggested replacement language.
The spreadsheet process appears to include several layers:
- The AI tool identifies text that may need adjustment.
- It proposes changes or rewrites.
- It shows how many words could be removed from individual regulations.
- It assigns a percentage figure for how noncompliant the regulations are.
- HUD staff are asked to review recommendations and justify objections.
The source article says it is not clear how those percentage figures are calculated. That uncertainty matters because the spreadsheet appears to turn legal and policy judgment into a numerical signal, while the basis for that signal remains unexplained in the reporting.
Human review is still part of the workflow
The source article does not describe a fully automated rulemaking process. Staffers from PIH are asked to review the AI recommendations and explain why they disagree with items they reject.
One HUD source described the setup this way: “It all sounds crazy—having AI recommend revisions to regulations,” while also saying they appreciated “how much they’re using real people to confirm and make changes.”
After PIH completes its review, recommendations are expected to go to the Office of the General Counsel for approval. That means the AI output is being routed through existing agency review channels, at least in the workflow described by the source.
Still, the scale and purpose of the project are significant. One HUD source said they were told the model is “being refined by our work to be used across the government.” The same source said they were told in a meeting attended by Sweet and Jacob Altik that the model would crawl through the Code of Federal Regulations, or eCFR.
Why HUD staff are concerned
The project sits inside a broader DOGE presence at HUD. DOGE representatives have been at the agency since February, when WIRED reported that two staffers received application-level access to some of the agency’s most critical and sensitive systems.
Sweet has also been given read access to HUD’s Public and Indian Housing Information Center and enterprise income verification systems, according to sources inside the agency. Those systems are connected to public housing and income verification data.
Earlier this month, US representative Maxine Waters, the top Democrat on the House Financial Services Committee, said DOGE had “infiltrated our nation’s housing agencies, stealing funding Congress provided to communities, illegally terminating staff, including in your districts, and accessing confidential data about people living in assisted housing, including sexual assault survivors.”
Some HUD sources also questioned the need for the regulation review. One source said rules already go through a “multi-year multi-stakeholder meatgrinder before any rule was ever created” under the Administrative Procedure Act. The source article notes that this law governs how agencies create regulations and allows for judicial oversight of agency actions.
Another HUD source questioned the description of Sweet’s role, saying: “A programmer and a quantitative data analyst are two very different things.”
The larger deregulation context
The HUD project is connected in the source article to a wider push for deregulation. Plans for large-scale deregulation across the US government were described in the Project 2025 policy document, which the Trump administration has effectively used as a playbook during its first 100 days in power.
That document, written by far-right figures, supports deregulation in areas including the environment, food and drug enforcement, and diversity, equity, and inclusion policies. Within that context, the AI review at HUD appears to be one operational example of a broader political goal: using software to scan government rules and identify language that can be reduced, replaced, or removed.
WIRED also linked Sweet to a public GitHub account, “CLSweet,” which created an application for tracking and analyzing federal regulations by showing how regulatory burden is distributed across government agencies. That application was last updated in March 2025, weeks before Sweet joined HUD.
Sweet has little online presence beyond that. The source article points to a short biography on East Edge Securities, an investment firm he founded in 2023 with two other University of Chicago students, and mentions claimed past work with private equity firms including Pertento Partners and Tenzing Global Investors. It also says he is listed as a board member of Paragon Global Investments and that the biography says he “will be joining Nexus Point Capital as a private equity summer analyst.”
For HUD, the immediate issue is not simply that AI is being used. It is that AI is being applied to the language of federal regulation, producing proposed rewrites and compliance judgments in a setting where legal authority, public housing policy, sensitive data, and political pressure all intersect.