GAO finds uneven safeguards in 10 Pentagon AI projects
The US Government Accountability Office found uneven application of the Pentagon’s Responsible AI guidance across 10 selected AI projects. Six respondents said their teams had identified operational risks, four said project requirements did not align with the department’s tenets, and three were unsure whether systems would be continuously monitored after deployment.
Bottom line — The Pentagon agreed to GAO recommendations on oversight and staffing, but both remain open.
Go deeper 5
-
The GAO sample cannot show how widely the findings apply across the department; officials reported 378 AI cases in February 2026 and said the actual total was probably higher.
-
The Responsible AI Division had filled three of its 10 authorised positions by February 2026.
-
The Responsible AI Working Council went without a meeting from late 2024 to August 2025; GAO said it lacked a charter setting out its responsibilities and meeting schedule.
-
GAO found that every component had taken at least one step to apply the governance tenet in its selected case.
-
A 2021 Pentagon memo set six Responsible AI tenets, including oversight, verification and real-time monitoring.