Skip to content
Neural Daily

PolicyNews

GAO: Pentagon's responsible AI office fills 3 of 10 posts

🇭🇺 Olvasd el magyarul →

The U.S. Government Accountability Office found that the Pentagon's Responsible AI Division had filled only three of ten authorized positions. Four AI project officials said their requirements are not aligned with responsible AI tenets.

By: Mikola Máté Patrik · · 2 min read

Aerial view of the Pentagon in Arlington, Virginia, with the Potomac River and the Washington Monument in the background
Image: U.S. Department of Defense · Public domain (US gov) · DoD photo by Master Sgt. Ken Hammond, U.S. Air Force, public domain via Wikimedia Commons · Source

What GAO reviewed

The GAO, the U.S. Congress's watchdog, published a report on October 14 on how the Department of Defense implements responsible AI (GAO). It surveyed ten DoD components and ten selected AI use cases, including target recognition and counter-camouflage systems, surveillance processing and tactical decision support.

What it found

  • Most components have taken at least initial steps to implement the six Responsible AI tenets set out in 2021.
  • Six of ten project officials said they had identified risks in using their systems in an operational environment.
  • Four said their project requirements were not aligned with Responsible AI tenets, and three did not know whether systems would be continuously monitored after deployment (
  • [object Object]
  • ).
  • The Responsible AI Working Council did not meet from late 2024 until August 2025 and still has no charter.
  • As of February 2026, only three of the CDAO Responsible AI Division's ten authorized positions were filled.

GAO made two recommendations: charter the council with defined roles and meeting frequency, and have the CDAO address the division's staffing. DoD concurred with both, but per NPR had implemented neither as of October 22. GAO also asks Congress to consider directing DoD to act on earlier recommendations on AI collaboration and workforce.

Why it matters

The Pentagon declared an AI-first strategy this year with the emphasis on speed. Its responsible AI tenets, however, live in a memo rather than a binding instruction. When oversight rests on three people, whether the tenets are followed comes down to each project team.

From a European angle, the EU AI Act explicitly excludes military uses from its scope. Oversight of military AI therefore depends on national institutions everywhere, and the GAO report shows how easily that slips in practice.

    Produced with the help of AI, reviewed by an editor. How we make our articles

    Reviewed and edited by the Neural Daily desk. About

    Related articles