A coalition of artificial intelligence policy groups is urging U.S. lawmakers to include specific guardrails on lethal autonomous weapons in the upcoming National Defense Authorization Act (NDAA), according to a report published Wednesday by The Hill.
The request, detailed in a letter sent to congressional leaders on June 3, 2026, argues that the Pentagon’s current policies on autonomous weapons systems are insufficient to prevent unintended escalation or civilian harm. The groups are pressing for language that would require meaningful human control over any weapon system that can select and engage targets without direct human intervention.
Signatories to the letter include prominent AI safety and policy organizations. They contend that as artificial intelligence technology advances rapidly, the U.S. military’s development of lethal autonomous weapons must be governed by clear, binding rules rather than internal guidelines that can be changed without congressional oversight.
Key demands in the letter
The coalition’s primary request is for the NDAA to mandate that all lethal autonomous weapon systems undergo rigorous testing and validation before deployment. The groups specifically call for a requirement that such systems be “reliably and predictably controllable” by human operators in all operational scenarios.
Another central demand involves transparency. The letter asks Congress to compel the Department of Defense to report publicly on the development, testing, and deployment of any weapon system with autonomous targeting capabilities. The groups argue that without such disclosure, lawmakers and the public cannot assess the risks these systems pose.
The policy groups also want the NDAA to prohibit the use of autonomous weapons in situations where civilian casualties are likely or where the system cannot distinguish between combatants and non-combatants with a high degree of reliability. This provision, they say, is essential to uphold international humanitarian law.
Congressional context
The letter arrives as the House and Senate Armed Services Committees are finalizing their versions of the fiscal year 2027 NDAA. The annual defense policy bill is considered a key vehicle for setting military AI policy, as it authorizes funding and establishes legal requirements for Pentagon programs.
Previous NDAA cycles have seen bipartisan interest in AI governance, but no comprehensive law has yet been enacted specifically for lethal autonomous weapons. The current push reflects growing concern among experts that the pace of AI development is outstripping the government’s ability to regulate it.
“The Department of Defense has its own directive on autonomous weapons, but it is a policy, not a law,” the letter states, according to The Hill’s summary. The groups argue that a policy can be changed by a single official, whereas a statutory requirement would provide lasting guardrails.
Opposing viewpoints
Some defense officials and industry representatives have argued that overly restrictive rules could hamper U.S. military competitiveness against rivals like China and Russia, who are also investing heavily in AI weapons. They caution that strict human-control mandates could slow down reaction times in combat scenarios where speed is critical.
Proponents of the guardrails counter that the risks of an AI arms race without rules are greater than the risks of slower development. They point to potential scenarios where autonomous systems could misinterpret data or be hacked, leading to unintended attacks that escalate into broader conflicts.
Several former military leaders have publicly supported the call for congressional action, stating that clear legal boundaries would actually strengthen U.S. leadership in setting global norms for AI in warfare.
The debate over lethal autonomous weapons is not new, but the current letter marks a coordinated effort to use the NDAA as the legislative vehicle for action. The groups are asking for language to be included in the final bill that is expected to be debated on the floor later this summer.
The next step to watch is whether the House or Senate Armed Services Committees include any of the proposed guardrails in their draft bills, which are expected to be released within weeks. If adopted, the provisions would represent the first major U.S. law specifically governing the use of AI in lethal military operations.
