Sponsors (2)
The bill strengthens privacy, human control, and oversight of Department of Defense AI/autonomy use—reducing catastrophic and privacy risks—but includes waiver authorities, broad exemptions, and provisions that could speed deployment and raise costs, creating tradeoffs between safety/oversight and operational flexibility and expense.
People in the United States gain protections because the bill bars AI-initiated nuclear launches and restricts AI-driven domestic surveillance, reducing risk of catastrophic or privacy-invasive actions.
Noncombatant civilians and military personnel benefit because human operators must retain the ability to intervene or deactivate autonomous systems, lowering the chance of unintended lethal engagements.
Military personnel and system operators get clearer standards: the Department of Defense must apply lifecycle testing, verification steps, and legal review before deploying AI/autonomous systems, improving consistency and accountability in procurement and operations.
Military personnel and civilians face heightened risk because waiver authority can let the DoD expedite use of AI/autonomy in real-world conflict or "compelling national interest" scenarios, potentially bypassing legal reviews and safeguards.
Military personnel, system developers, and civilians may be exposed to unchecked capabilities because many classes of systems (e.g., certain cyber, unarmed platforms, mines) are exempted from these oversight rules, leaving significant gaps.
Military personnel and civilians could see faster deployment of risky systems because the bill authorizes expanding autonomy when "practicable," which may accelerate fielding despite remaining operational risks.
Based on analysis of 2 sections of legislative text.
Requires DoD to adopt policy, risk categories, lifecycle testing, human oversight, and privacy protections for autonomous weapons and AI capabilities.
Official title: Amend title 10, United States Code, to establish a policy for the Department of Defense on maximizing autonomy and artificial intelligence systems, to establish requirements relating to Department review and verification of autonomous weapon systems, and artificial intelligence capabilities, and for other purposes.
Introduced June 8, 2026 by Christopher A. Coons · Last progress June 8, 2026
Requires the Department of Defense to adopt formal policy and rules for acquiring, developing, testing, deploying, and overseeing autonomous weapon systems and other AI capabilities. It directs the Secretary of Defense to both maximize appropriate use of autonomy and AI where practicable and ensure continuous human oversight, risk categorization, lifecycle testing, privacy protections, and the ability for humans to intervene or disable systems. Sets two risk-based levels for weapon systems (Level 1 = low risk if failure; Level 2 = moderate-to-high risk if failure) and two for AI capabilities (Level 1 = low risk from data or support failures; Level 2 = applies when Privacy Act or HIPAA-protected data are used or failure would severely impair mission). Applies to the Department of Defense and personnel using DoD autonomous weapons or AI capabilities and creates substantive operational and oversight requirements for those systems.