SweetREX: DOGE AI to Slash US Government Regulation

What if half of all federal regulations in the United States simply vanished? It sounds like a dream for some, a nightmare for others – but it's quickly becoming a very real possibility. Deep within the US government, an ambitious new **AI tool for policy** is being developed and deployed, poised to dismantle decades of regulatory frameworks at an unprecedented speed. The AI Revolution Sweeping Washington This isn't a future forecast; it's happening right now. On a recent video call, the Office of Management and Budget (OMB) unveiled "SweetREX Deregulation AI," an enigmatic new program designed to radically reshape how **federal regulations** are created, reviewed, and, more importantly, *eliminated*. This powerful **AI government deregulation** initiative isn't a standalone effort; it's a cornerstone of President Donald Trump’s "Unleashing Prosperity Through Deregulation" executive order and a core aim of **Project 2025**, the comprehensive playbook guiding a potential second Trump administration. The goal? To "promote prudent financial management and alleviate unnecessary regulatory burdens" by leveraging cutting-edge technology. Meet SweetREX: Your Digital Deregulation Dynamo Imagine an artificial intelligence capable of sifting through dense legal text, instantly identifying what's truly essential versus what's merely bureaucratic red tape. That's SweetREX. Still in active development, this sophisticated platform is engineered to pinpoint sections of regulations not explicitly required by statute. But it doesn't stop there. SweetREX then **turbocharges** the process, dramatically speeding up the adoption of updated, leaner regulations. The potential scale of this transformation is staggering. The so-called **Department of Government Efficiency (DOGE)**, a group central to these efforts, estimates that a colossal "50 percent of all federal regulations can be eliminated." This isn't just theory; it's a concrete target laid out in a July 1, 2025, PowerPoint presentation.

Image 1

Who is building this game-changing technology? SweetREX was born out of the collaborative efforts of DOGE associates, quietly working from within the Department of Housing and Urban Development (HUD). Their plan is clear: roll out this **regulatory burden reduction** powerhouse to agencies across the entire US government. Staffers from the Environmental Protection Agency (EPA), the Department of State, and the Federal Deposit Insurance Corporation (FDIC) were already on the call, getting a glimpse of this future. Inside the Engine Room: Who's Building the Future of Policy? At the heart of SweetREX's creation is Christopher Sweet, a DOGE affiliate and, until recently, a third-year student at the University of Chicago. Introduced to colleagues as a "special assistant," Sweet co-led the recent call and is identified as the primary developer behind the tool – hence its namesake. He revealed that **federal workers** are increasingly expected to leverage powerful AI platforms like those from Anthropic and OpenAI, with "a lot of the productivity boosts" stemming from tools built *around* these foundational technologies. For SweetREX specifically, the team is "primarily using the Google family of models, so primarily Gemini." Alongside Sweet on Wednesday's call was Scott Langmack, a senior adviser at HUD, also affiliated with DOGE. Langmack, whose LinkedIn profile lists him as the COO of proptech firm Kukun (a company "on a long-term mission to aggregate the hardest to find data"), is the front-facing advocate for SweetREX. While Sweet focused on development, Langmack is taking point on demoing the tool to various agencies, enthusiastically pitching its benefits. His bold claim? SweetREX can slash the time spent reviewing and proposing edited regulations from months down to a mere *few hours or days*. That's a promise of unprecedented **government efficiency**. Beyond the Code: AI's Impact on Public Input and Transparency SweetREX isn't just about cutting regulations; it's about reshaping the entire feedback loop. Its "decision tool" feature flags regulatory sections it deems "extraneous" based on relevant statutes. Government attorneys and policymakers can then review these AI-generated proposals and make final adjustments, with SweetREX even creating a draft of the **AI-altered regulation** for their consideration.

Image 2

But the tool's ambition extends further. Before new rules are adopted, agencies often publish an "Advance Notice of Proposed Rulemaking," inviting public comments. SweetREX is designed to group these public submissions into AI-determined "buckets," even categorizing the types of commenters – distinguishing between an individual citizen and a "sophisticated" corporate entity. This could profoundly alter how public feedback is processed and understood. However, this cutting-edge approach isn't without its challenges. A HUD worker previously told WIRED that combing through regulations flagged by the AI to explain *why they might still be needed* was "painful." Yet, during Wednesday's call, several participants were effusive in their praise, calling SweetREX "awesome" and "great." This stark contrast raises critical questions about the balance between automation and human oversight. The Unseen Hands and Unanswered Questions Despite the groundbreaking nature of this initiative, a veil of silence often surrounds it. Neither Christopher Sweet nor the OMB immediately responded to WIRED's requests for comment. HUD's press office offered only a terse "under review," and Google has yet to comment on their models' deployment. Scott Langmack, too, remained silent. And then there's Steve Davis. During the call, a person identified as Steve Davis asked if the software could be open-sourced and put on GitHub. This name resonates with controversy. Steve Davis is also Elon Musk's top lieutenant, who until recently served as a special government employee involved with running DOGE. He left government service to return as president of Musk’s Boring Company – a title he never abdicated even while serving in Washington. His departure reportedly left DOGE staffers uncomfortable, as he seemingly continued to exert influence despite no longer being a government employee. Davis also did not respond to WIRED's inquiries. As SweetREX prepares for wider rollout, critical questions linger. What are the long-term implications of AI-driven policymaking? How will transparency and accountability be maintained when powerful algorithms shape the very rules that govern us? And what impact will this **US government AI** push have on the lives of everyday citizens? The answers remain as elusive as the comments from its key players, leaving us to wonder: is this the future we asked for?

Image 3