Keeping pace with AI at the Department of State means accepting that the target is always moving.
The department has spent the past two years introducing employees to generative AI, training roughly 55,000 people and expanding StateChat, its internal enterprise generative AI platform, from an early experiment into a tool used across the department. But familiarity does not necessarily mean confidence. Employees still need to understand when AI is useful, when its answers should be questioned and how to use it safely.
During a NextGov/FCW webcast, Advancing AI from Answers to Actions, Erik Peterson, generative AI product owner for StateChat at the State Department’s Center for Analytics, and Stephen Wallace, public sector CTO at F5 and former CTO of the Defense Information Systems Agency, described an approach built around experimentation, user feedback and practical safeguards. Moreover, as AI’s role shifts from helping users answer simple questions,agencies need to understand where the technology fits, where it can fail and how to give employees room to use it responsibly.