Governments increasingly seek to implement AI, but adoption is lagging for several reasons, including lack of quality data and security concerns. And while some governments have been able to successfully implement AI, progress is fragmented.
Many states have made progress since the prior year’s first-ever report, doing so through executive orders, legislation, task forces and partnerships, the 2026 Government AI Landscape Assessment underlines. The report, released Friday, uses four stages of maturity to measure states’ progress in AI readiness: early, developing, established and advanced.
Readiness is about creating the capabilities for implementation. Piloting informs what will and will not work for scaled application in operations. Implementation entails the creation of measurable outcomes. Impact involves strong governance to fuel further innovation.
This year’s report highlights several key findings. Most states recognize AI as a strategic capability for public-sector operations. Nearly every state has launched some form of an AI pilot, often leveraging generative AI tools.
However, only a small number of states have integrated AI into enterprise-scale workflows across agencies. And few states have established the necessary evaluation mechanisms to determine the public value of AI deployments; only seven were classified as established in this area, per the report.
The states that are leading in government AI adoption have several things in common: strong executive leadership, cross-agency governance, sandbox environments for controlled experimentation, structured pilot programs, enterprise data infrastructure, and systems to measure outcomes. The report cites Maryland, New Jersey, North Carolina, Pennsylvania, Texas, Utah and Vermont as leading states.
