The U.S. Department of Government Efficiency (DOGE) is piloting a controversial artificial intelligence tool designed to eliminate a massive swath of federal regulations, to cut half of Washington’s regulatory mandates by the first anniversary of President Donald Trump’s second inauguration.
According to internal documents obtained by The Washington Post and four officials familiar with the effort, the tool — dubbed the “DOGE AI Deregulation Decision Tool” — has already begun analyzing nearly 200,000 regulations to determine which are legally unnecessary and potentially removable. A PowerPoint presentation dated July 1 outlines DOGE’s plans, estimating that about 100,000 regulations could be slashed, primarily through automated analysis and limited human oversight.
The tool has already been used to evaluate over 1,000 regulatory sections at the Department of Housing and Urban Development (HUD) in under two weeks. At the Consumer Financial Protection Bureau (CFPB), the AI system reportedly handled 100% of the deregulation effort. Three HUD employees confirmed that the AI was used to review large volumes of regulatory text and recommend changes or deletions.
The tool was developed by a team of engineers brought into government as part of Elon Musk’s DOGE project, two federal officials confirmed. These individuals spoke on condition of anonymity, as they were not authorized to discuss internal government activities.
Ambitious Goals, Controversial Methods
While conservatives have long argued that excessive federal regulation stifles economic growth, critics warn that regulations often exist to protect public health, safety, and the environment. The AI initiative, they say, raises legal and ethical concerns.
White House spokesman Harrison Fields said in a statement that “all options are being explored” to meet the president’s deregulation targets. While no single plan has been formally approved, he emphasized that the effort is “in its early stages” and being developed “in consultation with the White House.”
Fields also praised the DOGE team, calling them “the best and brightest,” and said they are leading an unprecedented effort to modernize government systems for efficiency and effectiveness.
A former DOGE legal chief, James Burnham, now managing partner at King Street Legal, said the team had worked hard to find creative legal and technological solutions to repeal as many rules as possible within the president’s term. “Using AI in this context is a logical strategy,” he added.
“Relaunch America” by January 2026
The PowerPoint lays out a plan for every federal agency to use the AI tool to generate a deregulation list within four weeks. Staff training is already underway, and agencies are expected to finalize their proposed cuts by September 1. The initiative culminates in a major announcement planned for January 20, 2026 — branded in the slides as “Relaunch America.”
The AI is expected to significantly reduce the time and effort typically required to repeal regulations. One estimate in the presentation claims it would take 3.6 million human hours to manually eliminate 100,000 regulations, but with AI, the job could be completed in just 36 hours of human oversight — a claimed 93% reduction in labor.
The presentation also claims the effort could save the government up to $3.3 trillion annually in compliance and enforcement costs. However, the methodology for those numbers is unclear.
Internal Pushback and Legal Doubts
Not all federal staff are on board. Three people involved in deregulatory discussions said many agency employees have resisted the effort, concerned about the tool’s accuracy and DOGE’s qualifications. The skepticism increased after Elon Musk's departure from the initiative, leading to diminished influence for DOGE at the agency level.
In one notable example, HUD staff reported that the AI tool made critical errors. In several instances, it misread legal texts, wrongly flagging regulations as being outside statutory authority. “It got the law wrong,” said one HUD employee involved in the review.
Still, the PowerPoint lists HUD and CFPB as successful pilot cases. An internal HUD email dated April 18 noted that the DOGE team was training staff to use the tool to review regulations for discrepancies with legal statutes. The same message indicated the team was working with legal experts to streamline the Administrative Procedure Act (APA) process and use AI for drafting notices and public comment responses.
A HUD spokesperson declined to provide specifics, stating only that the agency is exploring how to increase efficiency and that no final decisions have been made. “The intent is not to replace human expertise, but to support it,” they said.
CFPB did not respond to requests for comment.
Can AI Deregulate the Government?
The legal path to repealing regulations is governed by the APA, which requires agencies to provide notice, take public comment, and demonstrate that any rule change is reasonable and justified. Experts warn that these requirements cannot be bypassed — even with AI.
“There’s been a lot of flashy attempts to cut corners, but most don’t hold up in court,” said Nicholas Bagley, a law professor at the University of Michigan. “Even if a rule is repealed by AI, companies are unlikely to ignore it if it was not properly invalidated.”
DOGE appears aware of these concerns. The PowerPoint names four DOGE attorneys — James Burnham, Austin Raynor, Jacob Altik, and Ashley Boizelle — who purportedly vetted and approved the AI tool’s legal use. Three could not be reached for comment.
The document classifies federal regulations into three categories: 50% not required by law, 38% mandated by statute, and 12% necessary for agency function. The goal is to remove those deemed neither legally required nor operationally essential.
A Race Against the Clock
Despite Trump’s aggressive deregulation agenda, progress has been slower than some expected. The American Action Forum, a center-right think tank, estimates that regulatory cost savings this term have been modest compared to 2017. Without one major rule repeal, costs would have actually increased.
“They’re way behind the numbers from 2017,” said Doug Holtz-Eakin, the think tank’s president and former CBO director. “I expected them to dominate this effort again, but they haven’t.”
Part of the problem is staffing. While the Trump administration has emphasized shrinking the federal workforce, fewer personnel means fewer people available to handle deregulation efforts. “We’re expected to move quickly, but we don’t have enough people to draft the notices,” said one senior official.
Still, with the help of AI, DOGE hopes to meet its goals in time for Trump’s relaunch date in January 2026 — even if skepticism, legal barriers, and AI missteps make the road ahead uncertain.