We write to offer public comment on the National Science Foundation’s Request for Information on the Development of a 2025 National Artificial Intelligence Research and Development Strategic Plan (NSF-2025-OGC-0001) published on April 28, 2025. We are a group composed of policy researchers and data scientists employed by the Urban Institute, a nonprofit research and policy organization, who have researched local government deployment of AI and generative AI (genAI) tools. The views expressed here are our own and do not represent the Urban Institute, its trustees, or its funders.

An emerging narrative in AI policy is that regulation stifles American innovation and competitiveness. While overly cumbersome regulations can stall AI progress, our work finds that too little regulation can do the same.

Local governments play an essential role in national AI strategy as laboratories for AI testing across critical public services. While local governments have expressed excitement about the transformative potential of AI, they have largely held back from testing applications with the greatest potential impact due to the risk that the current lack of guidance, resources, and guardrails for safe AI experimentation could result in unintended public harm. This significant missed opportunity for innovation and impact hinders American efforts to be the unrivaled world leader in AI. To ensure that America continues to lead in AI development for transformative public impact, we encourage the inclusion or expansion of language referencing the following principles in the 2025 National AI R&D Strategic Plan: 

1. Further commitmentsto investments in AI and genAI testing, use, and evaluation by expanding upon language in Strategies 1, 3, 8, and 9 to explicitly include subnational governments in public–private partnerships to support investments in foundational and emerging AI technologies, use case research, and AI performance evaluation.
2. Center intergovernmental engagement and knowledge sharing throughout Strategies. Expanding upon language in Strategies 5, 8, and 9 to commit to intergovernmental partnerships and data sharing can accelerate knowledge sharing.
3. Expand references to community and stakeholder engagement to inform responsible and responsive AI research and use, particularly in service of Strategies 2 and 6 to bolster public trust in AI and ensure AI tools and performance evaluation criteria are responsive to public needs. 

We look forward to continued American leadership in AI research and development that spurs innovation while prioritizing security and responsible use.