PRIA Logo™
119-s-4707119th Congress

Responsible Artificial Intelligence Defense Act of 2026

Sponsored By: Senator Coons, Christopher A. [D-DE]

Introduced

Summary

Continuous human oversight of autonomous weapon systems and AI capabilities is the bill's core aim. This bill would create a Department of Defense policy to expand use of autonomy and artificial intelligence in weapons while requiring rigorous human judgment, legal review, testing, and privacy protections.

Show full summary
  • Service members and operators would have to exercise appropriate human judgment, continuously monitor AI-enabled weapons, and preserve ways to disengage or deactivate systems that behave unexpectedly. Verifications by the Under Secretary for Research and Engineering would remain valid for 3 years.
  • Civilians and U.S. persons gain specific limits on use. The bill bars AI from deciding a nuclear launch, bans autonomous lethal force without human judgment, and prohibits monitoring, profiling, or targeting people in the United States without a warrant.
  • Defense developers and acquisition programs must meet a formal verification, validation, testing, and cybersecurity regime and build safety and anti-tamper features. An Autonomy and AI Systems Working Group will advise on standards and DoD must submit annual reports on the policy through 2037.

Personalized for You

How does this bill affect your finances?

Personalize government policy and PRIA will tell you what this bill means for your household, plus every other piece of legislation we track. PRIA reads each provision against your financial profile to show you exactly what matters to your wallet.

Free to download

Bill Overview

Analyzed Economic Effects

5 provisions identified: 4 benefits, 0 costs, 1 mixed.

Bans on certain autonomous uses

This bill would bar the Department of Defense from using autonomy or AI to decide to launch a nuclear weapon. It would bar monitoring, tracking, profiling, or targeting people in the United States without a warrant based on probable cause or another constitutional basis. It would bar lethal force by fully autonomous weapons that lack appropriate human judgment. These prohibitions apply unless another statute says otherwise.

Testing, verification, and privacy checks

This bill would require layered review and rigorous testing before prototyping and again before fielding AI and autonomous weapon systems. Verifications would be valid for three years, and substantially similar variants would count as verified during that period. The Under Secretary for Research and Engineering or the Director for Operational Test and Evaluation would confirm safety, human-machine interfaces, cyber resilience, and iterative cyber testing. The Under Secretary could temporarily waive an updated legal review beyond one year during real conflict. For Level 2 AI that uses Privacy Act or HIPAA data or risks mission failure, the DoD must do a privacy impact assessment by the Director for Privacy, Civil Liberties and Transparency.

DoD AI policy and human oversight

This bill would tell the Department of Defense to use AI where practical while keeping continuous human oversight. The Secretary would have to make sure personnel exercise appropriate human judgment, monitor systems, and keep routine testing through a system's life. The Secretary would have to label non-exempt systems by risk level (Level 1 low risk, Level 2 higher risk including systems using Privacy Act or HIPAA data). The bill would also define key terms such as artificial intelligence, autonomy, and autonomous weapon system.

New DoD AI advisory working group

This bill would create an Autonomy and Artificial Intelligence Working Group at the Department of Defense. The group would advise technical reviewers and military leaders before formal development and before deployment. It would help decide when systems need senior-level approval and would develop safety standards for autonomy and AI.

Exemptions that narrow oversight

This bill would list ten categories of systems that are exempt from these AI and autonomy rules. Exemptions include operator‑supervised interceptors for time‑critical or saturation attacks, defenses for remotely piloted or autonomous vehicles and vessels, non‑lethal cyberspace capabilities, unarmed platforms, unguided or manually guided munitions, mines, and unexploded ordnance. The Secretary could exempt some systems not continuously monitored if the Secretary finds they are safer; congressional defense committees would have 30 days after notice to object.

Sponsors & CoSponsors

Sponsor

Coons, Christopher A. [D-DE]

DE • D

Cosponsors

  • Sen. Reed, Jack [D-RI]

    RI • D

    Sponsored 6/8/2026

Roll Call Votes

No roll call votes available for this bill.

View on Congress.gov
Back to Legislation