Elon Musk's Department of Government Efficiency is moving quickly on a custom generative AI chatbot for the US General Services Administration. The project, called GSAi, is being developed as part of President Donald Trump's AI-first agenda for modernizing federal work with advanced technology.
The effort is not just about giving employees another chatbot. According to two people familiar with the project, DOGE wants the tool to improve daily productivity for the GSA's roughly 12,000 employees and help analyze large volumes of contract and procurement data.
What GSAi is supposed to do
The General Services Administration has a central role inside the federal government. Its employees manage office buildings, contracts, and IT infrastructure across agencies, which makes it a natural target for tools that promise faster writing, analysis, and internal coordination.
GSAi is being discussed as a custom chatbot rather than a simple purchase of an existing commercial tool. One expected use is basic productivity support, such as helping workers draft memos faster. Another, more strategic use is contract analysis.
Thomas Shedd, a former Tesla employee who now runs Technology Transformation Services, referred to that work in a meeting on Wednesday. According to an audio recording obtained by WIRED, he said, “Another [project] I’m trying to work on is a centralized place for contracts so we can run analysis on them.”
Shedd also described the idea as connected to a larger effort to understand federal spending. “This is not new at all—this is something that’s been in motion before we started. The thing that’s different is potentially building that whole system in-house and building it very quickly. This goes back to this, ‘How do we understand how the government is spending money?’”
Why a custom chatbot matters
The decision to build GSAi followed discussions between the GSA and Google about Gemini, according to one person familiar with the project. The agency had hoped to use existing software such as Google Gemini, but later determined that it would not provide the level of data DOGE wanted.
That shift matters because it shows the project is about more than adopting the same chatbot tools now common in corporate America. ChatGPT and Gemini are already used in business settings for tasks such as writing emails and generating images. In the federal government, however, the adoption path has been more cautious.
Executive orders and guidance issued during the Biden administration generally told government staff to be careful with emerging technologies. President Donald Trump has taken a different approach, directing lieutenants to remove barriers to the US pursuit of “global AI dominance.”
DOGE has moved quickly in that environment. Reports from WIRED and other media say Musk's government efficiency team has recently pushed to bring more AI tools into federal work. At the same time, the broader changes have drawn heavy criticism from federal employees, labor unions, Democrats in Congress, and civil society groups, with some arguing that parts of the upheaval may be unconstitutional.
AI coding tools are part of the same push
GSAi is one piece of a larger AI agenda inside DOGE. On Monday, Shedd described deploying “AI coding agents” as one of the agency's top priorities, according to remarks described to WIRED.
These tools are designed to help engineers automatically generate, edit, and answer questions about software code. The goal is to raise productivity and reduce errors, but the source also describes a procurement and review process that has not been smooth.
One tool DOGE examined was Cursor, a coding assistant made by Anysphere, a fast-growing San Francisco startup. Anysphere's leading investors include Thrive Capital and Andreessen Horowitz, both of which have connections to Trump. Joshua Kushner, Thrive's managing partner, is the brother of Trump's son-in-law, Jared Kushner. Andreessen cofounder Marc Andreessen has said he's advising Trump on tech and energy policy.
A different person familiar with GSA technology purchases said the agency's IT team initially approved Cursor, then pulled it back for further review. DOGE is now pushing to install Microsoft's GitHub Copilot, described in the source as the world's most well-known coding assistant.
Cursor and the General Services Administration did not respond to requests for comment. Andreessen Horowitz and Thrive declined to comment. Google spokesperson Jose Castañeda declined to comment.
The review problem facing federal AI
The source article makes clear that federal AI interest did not begin with DOGE. In October 2023, then president Biden ordered the General Services Administration to prioritize security reviews for several types of AI tools, including chatbots and coding assistants.
By the end of his term, none had made it through even preliminary agency review processes, according to a former official familiar with them. As a result, no dedicated AI-assisted coding tools had received authorization under FedRAMP, a GSA program meant to centralize security reviews and reduce the burden on individual agencies.
Several agencies had still explored AI software. Transparency reports during Biden's term said the Commerce, Homeland Security, Interior, State, and Veterans Affairs departments were pursuing AI coding tools, including in some cases GitHub Copilot and Google's Gemini. GSA itself had been exploring three limited-purpose chatbots, including one for IT service requests.
The risk question remains central. Guidance from the personnel office issued under then president Biden said AI coding agent efficiency gains should be weighed against risks such as security vulnerabilities, costly errors, or malicious code. Federal regulations also require agencies to avoid even the appearance of a conflict of interest when choosing suppliers, and agencies are generally required by law to study possible cybersecurity risks before adopting new technology.
Fast adoption, unresolved questions
DOGE's AI work fits its broader effort to cut costs and speed up federal processes. The group has been working for the past few weeks to reduce spending across a government whose annual deficit has increased for the last three years. The Office of Personnel Management, described in the source as the government's HR department and as being stacked with Musk loyalists, has encouraged federal employees to resign if they cannot return to the office five days a week and commit to a culture of loyalty and excellence.
The same cost-efficiency logic appears in other departments. DOGE members at the Department of Education are reportedly using AI tools to analyze spending and programs, The Washington Post reported on Thursday. A department spokesperson said the focus is on finding cost efficiencies.
Yet the pace of adoption has already met limits. While DOGE has not publicly changed its stance, the team quietly halted the rollout of at least one generative AI tool this week, according to two people familiar with that project.
That leaves GSAi at the center of a larger federal test. The chatbot could help employees move faster and make contract data easier to analyze. But the same effort also raises questions about security review, supplier choice, and how much risk the government should accept when it decides that waiting is no longer an option.