Government AI Evaluation Case Stories Impact & Insights: Government AI Evaluation in Practice Understanding how and when government AI implementations are, and are not, working for the people they serve will help agencies focus their resources on fit-for-purpose use cases. Read our latest agency profiles below. To better serve the public, government agencies must consider the immediate and the long-term social impact of their AI interventions. Alongside technical tools like red teaming and model evaluation, agencies must integrate effective, achievable impact assessment into their AI programs. AI impact evaluation can be a low-burden cross-functional exercise, and it works best when collaborative and coordinated. This page brings together case studies and practitioner perspectives from state, local and federal government agencies, highlighting real-world examples of AI social impact evaluation in practice, featuring case studies, lessons learned and approaches to measuring outcomes. Case Stories Hear directly from other government practitioners and leaders doing this work. Demos, Not Memos How Colorado Is Rethinking AI Evaluation Read more A Source of Authenticity How the Library of Congress Approaches AI Read more Building the Conditions for Responsible AI How Maryland is Approaching AI Read more The complete government AI social impact evaluation toolkit, including guided navigation paths and evaluation tools, will launch soon.