Trends highlighted in the analysis include the rise of chief AI officers, investment in training programs, an evolving cybersecurity threat landscape, state-level policymaking, and secure sandbox environments for experimentation.
The Government AI Landscape Assessment explores AI readiness in three areas: leadership and governance, capacity building, and technical infrastructure and capabilities. The resource classifies states’ readiness levels in each of these areas under one of four categories: early, developing, established or advanced. The early classification includes states that have taken the initial steps in AI adoption, while the advanced classification recognizes states with sophisticated capabilities, frameworks and approaches.
States leading in readiness, according to this assessment, are Pennsylvania, New Jersey, and Utah, each of which received two “advanced” classifications and one “established” classification.
Each of these states has prioritized AI readiness. Pennsylvania has been testing and measuring AI for impact, and New Jersey is taking an economy-focused approach to AI and has been an early implementer of AI training. Utah has been an early AI adopter and even recently created an AI policy office that aims to answer societal AI questions.
Overall, in the category of leadership and governance, only three states were classified as advanced. Half, or 25, were classified as established; 16 as developing; and seven as early. Washington, D.C., was included as a state in this assessment. Utah and North Carolina were highlighted for their work in this area.
In technical infrastructure and capabilities, three states were classified as advanced, 16 as established, 23 as developing, and nine as early. Colorado and Minnesota were highlighted for their work in this.
“This analysis demonstrates what many of us know to be true: states are leading the way when it comes to adopting AI to make government more efficient and effective,” Jenn Thom, Code for America’s senior director of data science, said in a statement.
The assessment was created by reviewing public materials, AI-focused legislation and policy, guidance and reports, news coverage, and direct input.
Debate has arisen recently about whether AI policymaking should occur at the state or federal level, with the consensus largely being that both should have a role in regulation. With the removal of a provision to enact a moratorium on state-level AI regulation from the federal budget bill, states retain the authority to enact policy to guide responsible AI use.
