Governments are increasingly exploring ways to leverage AI, particularly generative AI (genAI), to bolster operational efficiency and improve delivery of public services. While local governments show interest in AI, Urban Institute research finds that they also face uncertainty about its responsible and secure use, express first-mover reluctance about specifically deploying genAI tools that may present risks and harms to constituents, and have limited resources for vetting and testing tools for policy applications.

Conveners, capacity builders, and resource aggregators, government support networks and technical assistance organizations are uniquely well-positioned to help local governments explore and adopt AI in a number of ways: 

• Many network and technical assistance organizations convene peer learning cohorts. Those that do can infuse AI-focused learning into their cohort curriculums, showcasing how AI can be applied to specific government functions or policy domains and at different levels of technical expertise across government staff.
• Government supporting organizations can help governments develop best practices for community engagement focused on public-facing AI and genAI use, outlining strategies for governments to solicit inclusive public input on data use permissions, priorities for AI pilots, and metrics to evaluate AI efficacy for a given task or policy goal.
• Organizations with more advanced technical expertise can directly test genAI tools in lower risk settings to offer governments guidance on their efficacy for use cases in specific policy domains, limitations, and compliance with government security requirements.
• Developing toolkits, practical guides, and repositories of AI resources can help local governments stay up to speed on and contribute to knowledge on technological and policy developments in the AI space.