DOGE Develops AI Chatbot to Scrutinize Federal Contracts

Compiled by the editorial desk with reference to reports from Wired and the Washington Post, along with public statements from government officials.

The Department of Government Efficiency (DOGE), led by billionaire Elon Musk, is developing a custom AI chatbot for the U.S. General Services Administration (GSA), according to a report from Wired. The tool is intended to analyze contracts and procurement data across the federal government, aiming to identify spending patterns and potential savings. The initiative, however, has raised concerns among IT and cybersecurity experts about the risks of entrusting sensitive data to a technology known for errors and data leaks.

DOGE, which has been moving rapidly through federal agencies, is part of a broader "AI-first strategy" that has gained traction since President Donald Trump revoked a 2023 executive order on AI safety. That order, issued by former President Joe Biden, had emphasized the need for "safe, secure, and trustworthy development and use of AI," particularly to prevent national security threats. Its revocation has cleared the way for a more permissive approach, often described as "move fast and break things."

Thomas Shedd, a former Tesla employee now heading the government's Technology Transformation Services, defended the AI initiative in an audio recording obtained by Wired. He described the creation of a "centralized place for contracts so we can run analysis on them" as "not new at all," framing it as a continuation of efforts to understand government spending. Shedd's remarks suggest that the AI chatbot is seen as a logical extension of existing data analysis efforts, rather than a radical departure.

Despite the administration's enthusiasm, the rollout has already encountered obstacles. According to Wired's sources, DOGE quietly halted the deployment of at least one generative AI tool this week. Additionally, the team had planned to use an AI coding assistant called Cursor but decided to pause that for "further review." Instead, DOGE has been encouraging workers to use Microsoft's GitHub Copilot, a more established tool in the coding space.

Security and Reliability Concerns

The move to integrate AI into federal operations has sparked alarm among experts who point to the technology's propensity for "hallucinations"—confidently stated but incorrect information—and its history of leaking sensitive data. These flaws make it a questionable choice for handling the vast amounts of confidential information held by the government. The GSA, which manages office buildings and IT infrastructure, is particularly sensitive, as its data could include details about federal employees, contractors, and operational security.

DOGE has already begun using AI tools within the Department of Education, feeding sensitive data into the software, as reported by the Washington Post. This early adoption, despite the known risks, underscores the administration's commitment to leveraging AI across agencies, even as experts warn of potential consequences.

Looking Ahead

The success of the GSA AI chatbot remains uncertain. While it could potentially speed up DOGE's efforts to reduce government spending, the technical challenges and security risks are significant. As the initiative progresses, the balance between innovation and caution will be a key factor in determining its outcome. For now, the cracks are already beginning to show, with the recent halts and tool changes indicating that the path forward is far from smooth.