State Government AI Landscape Analysis Part 4: Getting Ready for Government AI - What States Should Consider

The Government AI Landscape Assessment developed by Code for America provided us with the first comprehensive, nationwide snapshot of how ready state governments are to responsibly adopt and experiment with artificial intelligence. We evaluated leadership structures, policies, workforce capacity, infrastructure, and transparency practices across all 50 states. Through this work, we’ve seen the wide spectrum of readiness—and the common questions many leaders are now asking as they begin their own AI journey.

As states explore the path to responsible AI adoption, many ask: what are the first steps we should take? Our assessment revealed that readiness doesn’t require perfect systems—it starts with intentional choices that create space to learn and evolve.

Here are six key considerations for governments looking to build AI readiness across Leadership & Governance, AI Capacity Building, and Technical Infrastructure & Capabilities:

1. Start with Purpose: Rather than chasing hype, states should ask: what problems are we trying to solve? AI should align with public service goals, not the other way around. Early efforts should prioritize use cases that enhance equity, efficiency, or service access.
2. Empower Leadership Across Sectors: Strong AI readiness requires more than a single leader. Governors, CIOs, and agency heads must work together to champion AI, set clear direction, and ensure adoption is aligned with public values. Leadership at multiple levels helps ensure initiatives are supported, resourced, and integrated into agency operations.
3. Build Governance Early: Establishing basic structures—such as a task force, working group, or designated AI lead—signals commitment and helps coordinate early activities. Even temporary or advisory bodies can lay the groundwork for broader policies.
4. Create a Safe Space to Experiment: Pilots and sandbox environments help states test ideas in low-risk settings. Experimentation builds internal knowledge, reveals infrastructure needs, and uncovers where policy or data constraints may exist.
5. Invest in People, Not Just Tools: Readiness starts with human capacity. States should train civil servants, hire data talent, and engage community stakeholders. Equipping teams to evaluate and manage AI is as critical as the tools themselves.
6. Design for Trust and Transparency: States must ensure communities understand how AI is being used and can offer feedback. Even small steps—like publishing AI inventories or impact assessments—can improve accountability and build public confidence.

Examples of States Putting AI into Action

The six recommendations outlined above are more than theoretical—they’re grounded in real practice from states that are making progress:

• Start with Purpose: California’s wildfire response tools and North Carolina’s fraud detection systems each address core public service missions using AI, helping improve timeliness, targeting, and safety outcomes.
• Empower Leadership Across Sectors: Maryland established statutory AI governance backed by executive leadership and legislative engagement, ensuring AI policy coherence across state agencies.
• Build Governance Early: Utah’s Office of Artificial Intelligence Policy coordinates strategy and regulation, with an advisory board that ensures stakeholder engagement from day one.
• Create a Safe Space to Experiment: Connecticut and Utah use regulatory sandboxes and pilots to test AI with transparency and limited risk, while building lessons into long-term adoption.
• Invest in People, Not Just Tools: Massachusetts offers AI literacy courses for government employees and has partnered with MIT on workforce development. North Carolina embeds AI skills into onboarding and leadership training.
• Design for Trust and Transparency: Connecticut requires impact assessments for new AI tools and runs a statewide AI academy to build public understanding. Only 7 states today are in the "established" or "advanced" category for public engagement.

These data points from our assessment show that states can lead with creativity and care. Each of these efforts demonstrates a different facet of readiness—and together they reflect how governments can begin turning AI vision into sustained, ethical practice.  No single model fits all, but the point is that readiness requires deliberate steps. By embedding principles of ethics, equity, and public benefit from the start, governments can responsibly explore AI’s potential—and avoid its pitfalls.