Georgetown study finds gaps in measuring California and Florida’s AI readiness

California and Florida have built different strengths in AI, but neither has the comprehensive public-perception data needed for a new assessment of readiness. Georgetown researchers developed a framework that treats residents’ views, government purchasing and education as part of preparedness—not just technical capacity—and applied it to the two states.

The work, described by Georgetown’s McCourt School of Public Policy on October 21, comes after a burst of state activity. As of 2025, 47 states were considering AI-related legislation, and 34 had established a group or standing committee to study the technology.

Professor Lisa Singh led the team from Georgetown’s Massive Data Institute and Evidence for Justice Lab. Before building their own approach, the researchers examined state-level assessments from Code for America, the Council of State Governments and Rutgers University, alongside global indices from Oxford Insights, the International Monetary Fund and the United Nations.

Readiness beyond government and industry

Their criticism was that existing frameworks largely treated readiness as something governments and markets achieve. The people affected by AI received less attention. Public perceptions, equity and procurement were largely absent, the researchers found, even though governments often buy AI tools rather than build them.

The team’s response was a three-pillar framework, detailed in McCourt’s account of the research. It separates a state’s AI-specific capacity from the supporting systems and formal rules needed to put that capacity to use:

• Foundations covers public perceptions, AI literacy, innovation, implementation preparedness and actual use. It brings residents’ understanding and experience into the assessment alongside a state’s ability to adopt AI.
• Alignment covers the systems that help AI scale, including energy, infrastructure, data and education in science, technology, engineering and mathematics. These sit alongside, rather than inside, AI-specific capacity.
• Governance covers the formal structures shaping AI use: governing bodies, protections and incentives. Those incentives include grants and regulatory sandboxes, which the framework considers alongside the rules and institutions overseeing adoption.

Two states, different strengths

Applying the framework to California and Florida produced a comparison with strengths on both sides. Florida stood out in K–12 AI education. It adopted a statewide AI-specific Career and Technical Education program with the University of Florida, focused on technical proficiency and applied learning. Many high schools planned to adopt it.

California treats computer science as a standalone learning standard on par with math and science, but had not implemented a statewide AI-specific career program at Florida’s scale. Its stronger showing came in government readiness, according to the study.

California had tested models from major AI developers across several state agencies, issued purchasing guidelines for generative AI and offered training to government workers. The study found no equivalent testing environment or guidelines in Florida.

States need to understand how ready they are because having a baseline gives direction to where they need to invest next

Andrea Headley, associate professor and Evidence for Justice Lab faculty director

The missing baseline

The comparison also exposed a shared gap: comprehensive state-level data on public perceptions of AI was unavailable for either state. That is a missing measurement within the framework, not a finding that residents support or oppose adoption.

The researchers identify data collection as the first step toward measuring the framework’s components. They also call for sharing the resulting insights with the public to make AI adoption efforts more transparent. Their proposed next move is to establish the evidence needed for a baseline, then use it to guide investment.