Preparing the Public Workforce for AI

Agencies reported 3,611 AI use cases in 2025. What the government spends per employee preparing its workforce to use them is a number that does not exist, because it has never been tracked. This article makes one argument: readiness is a choice, drift is the current choice, and the gap between AI’s arrival and the workforce’s preparation will be paid in taxpayer dollars and degraded service. Against the private sector’s measured training benchmark, it presses for a published per-employee federal training figure, role-specific instruction in failure modes rather than features, and certification for the supervisors who govern high-impact AI. The government is deciding whether to pay wholesale now or retail later.