The US Senate Armed Services Committee has moved proposed military AI guardrails one step closer to becoming law by approving legislation that would regulate how the Pentagon develops, tests, and deploys artificial intelligence and autonomous weapons systems.
Included in the committee’s draft of the fiscal year 2027 National Defense Authorization Act, the measure would require meaningful human oversight of AI-enabled weapons, strengthen testing and accountability standards, and prohibit AI from making decisions to launch or detonate nuclear weapons.
AI Oversight Framework
If enacted, the legislation would require the Pentagon to ensure personnel exercise appropriate human judgment over the use of force, while commanders and operators retain ultimate responsibility for lethal decisions involving AI-enabled weapons.
Covered systems would need to support human supervision, intervention, manual override, and operational monitoring.
They would also be required to maintain records of target selection and decision logic while complying with international law, the law of armed conflict, applicable treaties, rules of engagement, and Pentagon policy.
The bill would also codify the Pentagon’s existing review process, requiring verification, validation, testing, and evaluation before deployment.
Under the proposal, the Defense Department’s Director of Operational Test and Evaluation would assess whether covered systems can operate reliably in realistic environments against adaptive adversaries.
An incident repository would also track system failures, unintended behavior, and near misses to strengthen oversight and future improvements.
Broader Military AI Initiatives
Beyond the new guardrails, the draft legislation also promotes wider adoption of AI across the military, including expanded use of generative AI and autonomous software agents.
It also calls for establishing a Robotics and Autonomous Systems Command to oversee training, interoperability, doctrine development, and the operational employment of autonomous systems.
