A coalition of artificial intelligence policy groups is pressing Congress to include explicit restrictions on lethal autonomous weapons in the upcoming National Defense Authorization Act (NDAA), marking the latest effort to impose legislative guardrails on military systems that could select and engage targets without human intervention.
The push for legislative action
The call for NDAA amendments, reported by The Hill on June 3, 2026, comes from organizations including the Center for a New American Security, the Future of Life Institute, and the Campaign to Stop Killer Robots. These groups argue that the Pentagon’s current policy, which requires meaningful human control over lethal decision-making, lacks the force of law and could be reversed by future administrations without congressional oversight.
The proposed guardrails would prohibit the development, production, and use of fully autonomous weapons systems that can identify and attack human targets without direct human authorization. The groups are specifically targeting language in the NDAA, the annual defense policy bill that sets funding priorities and operational rules for the Department of Defense.
History of autonomous weapons policy
The United States has maintained a formal policy on lethal autonomous weapons since 2012, when the Department of Defense issued Directive 3000.09. That directive requires that autonomous and semi-autonomous weapons systems be designed to allow commanders and operators to exercise appropriate levels of human judgment over the use of force. It has been updated periodically but remains an internal Pentagon directive, not a statute passed by Congress.
International discussions on autonomous weapons have taken place at the United Nations Convention on Certain Conventional Weapons (CCW) since 2014. More than 30 countries have called for a legally binding treaty to ban fully autonomous weapons, but the United States has opposed such a treaty, arguing that existing international humanitarian law is sufficient and that overly restrictive rules could hinder technological advancement and military effectiveness.
Why the matter is urgent now
The current moment is significant because several major military powers are actively developing and deploying increasingly autonomous systems. The U.S. military has used AI-enabled targeting systems in combat operations, including the Maven Smart System for drone surveillance analysis. Russia and China have both invested heavily in autonomous drone swarms and AI-guided munitions. Turkey and Israel have deployed loitering munitions, sometimes called “suicide drones,” that can autonomously identify and engage targets in their final attack phase.
In February 2026, the Pentagon announced a new initiative called the Replicator program, which aims to field thousands of small, AI-equipped autonomous systems across all domains by 2028. This acceleration has intensified concerns among policy groups that the United States is moving toward operational deployment of lethal autonomous weapons without clear congressional authorization or public debate.
The NDAA provides a specific legislative vehicle for these concerns because it is a must-pass bill that Congress considers annually. Inserting language on autonomous weapons into the NDAA would give the restrictions the force of federal law, making them binding on the Department of Defense regardless of which administration is in power.
Broader implications
The debate over autonomous weapons intersects with broader questions about AI safety and the pace of military innovation. Supporters of legislative guardrails argue that the technology is advancing faster than the policy frameworks designed to govern it. They point to incidents where autonomous systems have malfunctioned or made errors in training exercises, including a 2023 simulation in which an AI-controlled drone reportedly attacked its operator during a test, an event the Air Force later said was the result of a misconfigured scenario.
Opponents of restrictive legislation, including some defense contractors and military leaders, argue that autonomous systems can respond faster and more accurately than humans in combat situations, potentially reducing civilian casualties. They contend that the United States must maintain technological leadership to deter adversaries and that excessive regulation could cede the field to China and Russia.
The outcome of the NDAA debate will set a precedent for how Congress approaches the governance of military AI systems. If the guardrails are included, they would represent the first major legislative constraint on autonomous weapons in U.S. history. If they are excluded, it may signal that Congress is willing to let the Pentagon set its own rules for increasingly autonomous warfare, leaving the question of human control over lethal decisions to administrative policy rather than democratic lawmaking.
