Lawmakers in the U.S. Senate are preparing legislation to restrict how the military uses artificial intelligence, following a high-profile dispute between Anthropic, a leading AI developer, and the Department of Defense. The move aims to codify ethical boundaries around autonomous weapons and mass surveillance, ensuring human oversight in life-or-death decisions. This comes after the Trump administration blacklisted Anthropic for refusing to allow unrestricted military application of its AI models, a decision Anthropic is now challenging in court.
The Core of the Conflict
Anthropic publicly resisted a Pentagon deal that would have permitted the use of its AI for fully autonomous weapons systems and broad domestic surveillance. This stance contrasts sharply with OpenAI, a competitor that agreed to the terms. The administration’s subsequent blacklisting of Anthropic has sparked legal action and now, legislative intervention.
The underlying issue is control: Who decides when and how AI is used in critical applications, especially those involving lethal force or mass data collection? Without clear rules, AI development could accelerate unchecked military use, raising serious civil liberties concerns.
Bipartisan Efforts to Set Guardrails
Senator Adam Schiff (D-CA) is leading the effort to draft a bill that would legally enforce limits on AI’s military applications. He emphasized the need for human involvement in any decision with life-or-death consequences, stating, “We don’t want to delegate that kind of responsibility over life and death to an algorithm.”
Senator Elissa Slotkin (D-MI) has already introduced the AI Guardrails Act, which would further restrict the Department of Defense from using AI for mass surveillance of Americans or autonomous lethal weapons without human oversight. The bill allows exceptions only under “extraordinary circumstances,” requiring Congressional notification.
The Role of AI in Warfare: A Delicate Balance
While emphasizing human oversight, lawmakers acknowledge AI’s potential military benefits. AI can process information faster than humans, providing a tactical advantage on the battlefield. However, the risk of errors—distinguishing between civilians and combatants, for example—remains significant.
The debate centers on how to harness AI’s speed and efficiency without relinquishing control over life-or-death decisions. The proposed legislation seeks to strike that balance.
Political Challenges and Next Steps
Passing such legislation will be challenging given the current political climate. With Democrats holding a narrow majority, bipartisan support will be crucial. The timing is also unfavorable, as midterms approach and legislative momentum slows. Lawmakers are considering attaching the bill to the National Defense Authorization Act (NDAA) to increase its chances of passage.
Despite political hurdles, Schiff believes there is broad public support for AI limitations. He acknowledges potential resistance from some colleagues who may view the bill as criticism of the administration, but remains optimistic about securing bipartisan backing.
“There’s certainly bipartisan support in the public for these kinds of limitations,” Schiff said.
The conflict with Anthropic has exposed vulnerabilities in current AI governance and accelerated the push for legal safeguards. Whether Congress can act swiftly to prevent unchecked military application of AI remains to be seen, but the debate underscores the urgent need for ethical clarity in this rapidly evolving field.
