Code for America Maps State Progress on Government AI Adoption
Code for America released its second annual assessment of state AI readiness on May 1, showing that states are moving beyond experimental pilots to embed AI into daily government operations. The nonprofit evaluated all 50 states across four stages: readiness, piloting, implementation, and impact measurement.
Seven states stand out for their progress: Maryland, New Jersey, North Carolina, Pennsylvania, Texas, Utah, and Vermont. These states have established executive leadership for AI, built cross-agency governance structures, and created controlled pilot programs with clear evaluation timelines.
What the Assessment Found
More states are advancing from initial pilots to active experimentation. However, most remain in early stages when it comes to operationalizing AI systems and measuring long-term outcomes.
The report examined how states use AI specifically for benefits administration-helping eligible residents find, apply for, and receive public assistance. This focus reflects where government AI work is actually happening.
How Leading States Are Using AI
Maryland partnered with Anthropic to deploy an AI agent that walks residents through the benefits application process.
New Jersey expanded its AI Assistant with specialized tools for document validation, public-facing content generation, and analysis of resident feedback at scale.
North Carolina uses generative AI to summarize documents, freeing caseworkers from routine reading tasks.
Pennsylvania deployed an AI tool to scan documents for legibility during benefit applications, reducing administrative work for caseworkers.
Texas published a formal AI governance framework and is improving data quality across agencies as it scales beyond pilots.
Utah launched a regulated pilot program using an AI chatbot to accelerate prescription renewals, allowing doctors to focus on patient care.
Vermont created a public inventory of all AI tools currently in use by state agencies.
What This Means for Government Workers
The assessment shows that government agencies are not waiting for perfect conditions before adopting AI. States are moving forward with clear governance structures and measurable goals.
Code for America conducted the research by reviewing executive orders, legislation, agency reports, and media coverage. Each state had the opportunity to review findings before publication.
In 2025, Code for America worked in 27 states and Washington, D.C. to help 7 million people access $22 billion in benefits across safety net, criminal justice, and tax systems.
For government professionals overseeing or implementing AI initiatives, the report provides a benchmark for where your state stands and what peer states are doing. Learn more at Code for America's website.
For those responsible for AI policy or governance decisions, the AI Learning Path for Policy Makers covers the governance frameworks and decision-making approaches that inform these state-level efforts.
Your membership also unlocks:
