Senate Lawmaker Presses DoD and Tech Firms to Disclose AI Contract Terms

Senator Elizabeth Warren is demanding that the Department of Defense and seven major technology companies disclose the full terms of agreements related to the deployment of artificial intelligence on classified networks. The inquiry focuses on whether current contractual guardrails are sufficient to prevent the misuse of AI for domestic surveillance or the development of autonomous weapon systems. This oversight push highlights increasing legislative scrutiny over how the defense industrial base integrates emerging technologies into sensitive military operations.

In May, the Department of Defense (DoD) established agreements with SpaceX, OpenAI, Google, NVIDIA, Microsoft, Amazon Web Services, Oracle, and Reflection AI to integrate AI capabilities into Impact Level 6 (IL6) and Impact Level 7 (IL7) network environments. IL6 is designated for processing information classified up to the secret level, while IL7 supports highly restricted data. Senator Elizabeth Warren (D-Mass.) has issued letters to Defense Secretary Pete Hegseth and the involved companies, arguing that the lack of public or congressional access to these contracts makes it impossible to assess existing safeguards. She noted that the DoD has provided few details beyond stating the technology is for "lawful operational use," a term critics argue is too broad to effectively constrain government actions.

The inquiry specifically highlights concerns regarding Reflection AI, a startup backed by the venture capital firm 1789 Capital, where Donald Trump Jr. is a partner. Warren has asked for clarification on the criteria used to award the contract to the startup, which has yet to release a public AI model, and whether any department officials communicated with Trump Jr. regarding the deal. Additionally, the Senator pointed to the DoD's treatment of Anthropic, which was designated a "supply chain risk" after refusing to remove contractual limitations on AI usage. Reports suggest the Pentagon may have used negotiations with other firms to pressure Anthropic into dropping safety concerns related to autonomous weapons and domestic surveillance.

This push for transparency aligns with broader legislative efforts to regulate military AI. Representatives and Senators, including Elissa Slotkin (D-Mich.) and Kirsten Gillibrand (D-N.Y.), have introduced bills to prohibit the use of autonomous weapons for lethal strikes without human authorization and to ban AI-driven mass surveillance or nuclear launches. The Senate Armed Services Committee has already integrated similar restrictive language into its version of the 2027 defense policy bill. Warren has requested that the DoD and the contracted tech firms provide unclassified responses regarding their contractual guardrails by July 20, emphasizing that vague language like "applicable law" may not prevent the technology from being used in ways that endanger civil liberties.