The Pentagon says AI use has surged to 1.5 million personnel as of this month, yet adoption remains below 50% of its workforce—and questions linger over whether the actual work is better.

What happened: Pentagon Chief Technology Officer Emil Michael disclosed that 1.5 million personnel are now using commercial AI tools, up from 80,000 in December 2025. That represents roughly 43% of the Department of Defense's 3.5 million employees. The surge has been fueled by partnerships announced in May with SpaceX, OpenAI, Google, NVIDIA, Reflection, Microsoft, Amazon Web Services, and Oracle, as well as discounted rates from tech companies via the General Services Administration.

Why it matters: The Pentagon credits AI with reducing the time to produce mandatory Congressional reports from 200 hours of staffing time to five hours—a major time-saving claim that aligns with the initial promise of efficiency gains. However, broader analysis by Brookings Institute found adoption remains concentrated in five large federal agencies, with 11 small agencies accounting for just 2% of total AI inventory use. Workforce capacity constraints, risk-averse culture, and low public trust in AI systems are slowing wider deployment.

What to watch: Only 9% of workers surveyed trust AI for complex tasks, and more than half bypass AI tools to do work manually instead. A Government Accountability Office report in March warned that without clear privacy guidance from the Office of Management and Budget, agencies risk exposing sensitive information through increased AI data access.