Elon Musk’s so-called Department of Government Efficiency (DOGE) operates on a core underlying assumption: The United States should be run like a startup. So far, that has mostly meant chaotic firings and an eagerness to steamroll regulations. But no pitch deck in 2025 is complete without an overdose of artificial intelligence, and DOGE is no different.

AI itself doesn’t reflexively deserve pitchforks. It has genuine uses and can create genuine efficiencies. It is not inherently untoward to introduce AI into a workflow, especially if you’re aware of and able to manage around its limitations. It’s not clear, though, that DOGE has embraced any of that nuance. If you have a hammer, everything looks like a nail; if you have the most access to the most sensitive data in the country, everything looks like an input.

Wherever DOGE has gone, AI has been in tow. Given the opacity of the organization, a lot remains unknown about how exactly it’s being used and where. But two revelations this week show just how extensive—and potentially misguided—DOGE’s AI aspirations are.

At the Department of Housing and Urban Development, a college undergrad has been tasked with using AI to find where HUD regulations may go beyond the strictest interpretation of underlying laws. (Agencies have traditionally had broad interpretive authority when legislation is vague, although the Supreme Court recently shifted that power to the judicial branch.) This is a task that actually makes some sense for AI, which can synthesize information from large documents far faster than a human could. There’s some risk of hallucination—more specifically, of the model spitting out citations that do not in fact exist—but a human needs to approve these recommendations regardless. This is, on one level, what generative AI is actually pretty good at right now: doing tedious work in a systematic way.

There’s something pernicious, though, in asking an AI model to help dismantle the administrative state. (Beyond the fact of it; your mileage will vary there depending on whether you think low-income housing is a societal good or you’re more of a Not in Any Backyard type.) AI doesn’t actually “know” anything about regulations or whether or not they comport with the strictest possible reading of statutes, something that even highly experienced lawyers will disagree on. It needs to be fed a prompt detailing what to look for, which means you can not only work the refs but write the rulebook for them. It is also exceptionally eager to please, to the point that it will confidently make stuff up rather than decline to respond.

If nothing else, it’s the shortest path to a maximalist gutting of a major agency’s authority, with the chance of scattered bullshit thrown in for good measure.

At least it’s an understandable use case. The same can’t be said for another AI effort associated with DOGE. As WIRED reported Friday, an early DOGE recruiter is once again looking for engineers, this time to “design benchmarks and deploy AI agents across live workflows in federal agencies.” His aim is to eliminate tens of thousands of government positions, replacing them with agentic AI and “freeing up” workers for ostensibly “higher impact” duties.

Here the issue is more clear-cut, even if you think the government should by and large be operated by robots. AI agents are still in the early stages; they’re not nearly cut out for this. They may not ever be. It’s like asking a toddler to operate heavy machinery.

DOGE didn’t introduce AI to the US government. In some cases, it has accelerated or revived AI programs that predate it. The General Services Administration had already been working on an internal chatbot for months; DOGE just put the deployment timeline on ludicrous speed. The Defense Department designed software to help automate reductions-in-force decades ago; DOGE engineers have updated AutoRIF for their own ends. (The Social Security Administration has recently introduced a pre-DOGE chatbot as well, which is worth a mention here if only to refer you to the regrettable training video.)

Even those preexisting projects, though, speak to the concerns around DOGE’s use of AI. The problem isn’t artificial intelligence in and of itself. It’s the full-throttle deployment in contexts where mistakes can have devastating consequences. It’s the lack of clarity around what data is being fed where and with what safeguards.

AI is neither a bogeyman nor a panacea. It’s good at some things and bad at others. But DOGE is using it as an imperfect means to destructive ends. It’s prompting its way toward a hollowed-out US government, essential functions of which will almost inevitably have to be assumed by—surprise!—connected Silicon Valley contractors.

The Chatroom

What are some ways AI could actually improve government efficiency? Leave a comment below or send your thoughts to mail@wired.com.

WIRED Reads

• A DOGE Recruiter Is Staffing a Project to Deploy AI Agents Across the US Government: A startup founder told a Palantir alumni Slack group that AI agents could do the work of tens of thousands of government employees. He was met with emojis of clowns and a man licking a boot.

• DOGE Put a College Student in Charge of Using AI to Rewrite Regulations: A DOGE operative has been tasked with using AI to propose rewrites to the Department of Housing and Urban Development’s regulations—an effort sources are told will roll out across government.

• Trump’s Quest for Crypto Riches Is a Constitutional Scandal Waiting to Happen: The largest investors in Donald Trump’s crypto coin have been promised a private audience with the president. The arrangement could put Trump at odds with the Constitution and add fuel to calls for his impeachment, experts say.

Want more? Subscribe now for unlimited access to WIRED.

What Else We’re Reading

🔗 Senior State Department official sought internal communications with journalists, European officials, and Trump critics: What if we did the “Twitter Files” but it’s by the government? (MIT Tech Review)

🔗 White House launches Drudge-style website to promote Trump: What if we did the Drudge Report, but it’s from the White House? (Axios)

🔗 A White House Briefing Straight From North Korea: What if we did the White House press corps, but it’s all alt right? (The Atlantic)

The Download

This week, our flagship Uncanny Valley podcast talked about the dangerous decline in US vaccination rates and Robert F. Kennedy Jr.’s role in shaping US attitudes toward jabs. Listen now.

This is an edition of the WIRED Politics Lab newsletter. Read previous newsletters here.