Lawmakers in the Senate are introducing new legislation aimed at curbing how the U.S. military employs artificial intelligence, especially to prevent AI-driven nuclear weapon launches or domestic surveillance. This move coincides with discussions surrounding the annual National Defense Authorization Act, where AI ethics and control are becoming hotly contested issues.
Senator Kirsten Gillibrand has introduced the Secure and Accountable Military AI Act, which would bar AI from autonomously executing nuclear strikes, conducting surveillance on Americans, or developing and deploying fully autonomous weapons systems. The bill mandates that any high-stakes military actions, such as lethal targeting, detention decisions, cyber operations, and nuclear command controls, require meaningful human involvement and senior official approval.
Another senator, Elissa Slotkin, is preparing a complementary amendment that echoes these restrictions, notably prohibiting the Pentagon from using AI in launching nuclear weapons or in autonomous killing absent human oversight. These legislative efforts respond to growing concerns about the ethical and safety implications of handing critical defense decisions over to AI systems that might misinterpret data or act unpredictably.
The debate extends beyond military applications to issues of privacy and the potential consequences of large-scale AI integration within defense infrastructure. Critics warn that rapid adoption without adequate safeguards could lead to disastrous outcomes, endangering both national security and civilian safety.
Gillibrand has emphasized that the Pentagon is advancing AI capabilities without appropriate guardrails, which could undermine the security of service members and the general public. This initiative reflects a broader caution within government circles about unchecked AI deployment in sensitive areas.