Classified estimates show the NSA is paying billions to test AI models
The National Security Agency told lawmakers it is spending billions of dollars in taxpayer funds this year evaluating and testing advanced artificial intelligence models, according to two sources familiar with classified intelligence estimates. That price tag is significantly greater than previously known, and it has led lawmakers to believe a more comprehensive AI regulatory system could cost the government tens of billions of dollars per year.
The spending stems from the NSA's Artificial Intelligence Security Center, which began testing frontier models to identify potential national security vulnerabilities after a string of high-profile hacking and security incidents. President Donald Trump has mostly resisted calls for greater federal oversight of AI, rejecting regulatory proposals from members of Congress and the frontier labs aimed at imposing new safeguards.
The high cost is bringing new urgency to the debate over the government's role in reviewing AI models built by some of the world's most powerful and resourceful companies, and may intensify calls for frontier labs to shoulder the financial burden of that review — especially as spending is set to balloon as the technology advances. Prior proposals carried far lower estimates: the Congressional Budget Office estimated the bipartisan AI Security and Innovation Act, which would establish a center on AI risks and "facilitate the mitigation of those risks," would cost roughly $20 million per year, while a separate House bill creating a new AI reporting and tracking system would cost $36 million total over five years.
Alternatives are circulating. Billionaire Elon Musk has suggested the frontier labs should review each other's models before release without government oversight, and on Thursday The Information reported that Google, OpenAI and Anthropic are jointly working on a plan to create their own AI safety-focused "standards body" — a model some AI safety experts reject as effectively letting firms police themselves. Another option is for the government to levy a tax on AI companies to run these safety operations, ensuring both independence and that the financial burden does not fall primarily on taxpayers; Anthropic and OpenAI have suggested they want greater federal oversight and may be open to paying for it.
NSA's AI testing efforts are currently funded by classified portions of the federal national security budget, the two sources said, and the exact amount spent so far is not clear. The Pentagon declined to comment on what the NSA is spending on AI, with a spokesperson saying, "For security reasons, the Department does not discuss the technical architecture or resource allocation for its AI tools."