TL;DR:
• DOGE deploys SweetREX AI to identify "extraneous" federal regulations for elimination
• University of Chicago student Christopher Sweet leads development using Google's Gemini models
• Tool aims to cut "50 percent of all federal regulations" according to internal DOGE estimates
• EPA, State Department, and FDIC among agencies now testing the deregulation platform
The Trump administration just weaponized artificial intelligence for mass deregulation. A new AI tool called SweetREX, developed by a University of Chicago undergraduate turned DOGE affiliate, is now scanning federal regulations and flagging rules for elimination across multiple government agencies. The tool represents the most aggressive use of AI for systematic government downsizing in US history.
The Department of Government Efficiency just unveiled its secret weapon for dismantling federal oversight: an AI system that promises to slash government regulations faster than any human bureaucrat ever could. During a Wednesday video call hosted by the Office of Management and Budget, officials demonstrated SweetREX Deregulation AI, a tool that automatically scans federal rules and flags sections it deems unnecessary for elimination.
The system's developer is Christopher Sweet, a University of Chicago undergraduate who took leave to become a DOGE affiliate. Named after its creator, SweetREX (Sweet Regulation EXaminer) runs primarily on Google's Gemini models and represents the Trump administration's most ambitious attempt to automate government downsizing. "A lot of the productivity boosts will come from the tools that are built around these platforms," Sweet told government officials on the call, according to WIRED's reporting.
The timing couldn't be more strategic. President Trump's "Unleashing Prosperity Through Deregulation" executive order explicitly calls for eliminating "unnecessary regulatory burdens," while internal DOGE presentations obtained by The Washington Post estimate that "50 percent of all federal regulations can be eliminated." SweetREX appears designed to make those cuts systematic and scalable.
Government agencies are already lining up to test the platform. The call included staffers from the Environmental Protection Agency, Department of State, and Federal Deposit Insurance Corporation, among others. Scott Langmack, a DOGE-affiliated senior adviser at the Department of Housing and Urban Development where SweetREX was initially developed, claims the tool can reduce regulation review times "from months to just a few hours or days."
The technical capabilities reveal the scope of DOGE's ambitions. SweetREX's "decision tool" automatically flags regulatory sections it considers extraneous based on statutory requirements, then generates draft revisions for human review. The system also processes public comments on proposed rule changes, sorting them into "AI-determined buckets" and categorizing commenters as individuals versus "sophisticated" corporate entities.
But the human cost is becoming apparent. A HUD worker told WIRED last month that reviewing AI-flagged regulations to justify their necessity was "painful." The comment underscores how SweetREX shifts the burden of proof – regulations must now justify their existence to an algorithm trained on efficiency metrics rather than public safety considerations.
The platform's rollout reflects DOGE's broader strategy of embedding tech industry approaches within government operations. Sweet told colleagues that tools from Anthropic and OpenAI will become standard for federal workers, positioning SweetREX as just the beginning of AI-driven government transformation.
Even former DOGE leadership remains involved. Steve Davis, identified as Elon Musk's top lieutenant who officially left his government role to return to the Boring Company, appeared on the call asking whether SweetREX could be open-sourced on GitHub. His continued involvement despite leaving government has reportedly made current DOGE staffers uncomfortable, according to Politico.
The implications extend far beyond government efficiency. Kukun, the proptech company where Langmack serves as COO, describes itself as being "on a long-term mission to aggregate the hardest to find data." His dual role raises questions about how private sector interests might influence which regulations get flagged for elimination, particularly in housing and financial sectors where Kukun operates.
For now, SweetREX remains in development, but its rapid deployment across multiple agencies signals DOGE's intent to institutionalize AI-driven deregulation before the current administration's mandate expires. The question isn't whether artificial intelligence can streamline government operations – it's whether delegating regulatory decisions to algorithms serves the public interest or simply makes dismantling oversight more efficient.
SweetREX represents more than just another government efficiency tool – it's the automation of ideological policy implementation. By training artificial intelligence to identify "unnecessary" regulations, DOGE has created a scalable mechanism for systematic deregulation that operates faster than traditional oversight can respond. Whether this approach delivers the promised economic benefits or simply accelerates the dismantling of consumer and environmental protections will likely define the legacy of AI's first major deployment in American governance.
