A national blueprint for responsible AI in government

The debate over whether government should adopt artificial intelligence is over. There is no question at this point that AI has a role to play — now, a new journey has begun. 

States are racing to adopt AI amid new federal policies that will increase the costs of delivering critical assistance like food stamps and health insurance to millions of Americans. 

Rather than blindly chasing the latest AI tools, as the current Administration is doing, state leaders are focused on a more meaningful challenge: ensuring AI is adopted quickly and responsibly, to meet the needs of their residents.

That’s because they recognize that AI leadership means prioritizing transparency, safety and trust over hype. Ultimately, states are making sure that AI actually works for people, not against them.

In red and blue states alike, AI-enabled solutions are creating a more efficient and effective government. AI is being used to streamline administrative duties, answer common questions, and save public employees valuable time.

In the process, AI is opening the door to more empathetic interactions — something those of us who have had to renew an expired drivers license, apply for unemployment while in between gigs, or submit a complicated state tax return can appreciate.

Of course, some states are further down this road than others. According to Code for America’s recent Landscape Assessment, three states have emerged as national leaders: Pennsylvania, New Jersey and Utah. These three states are leading the pack when it comes to balancing innovation and responsible AI development.

What distinguishes these states from the rest?

Each is advancing a bold blueprint for AI — one that goes beyond simply rolling out the latest new tools and instead is offering a clear, replicable roadmap for the future. 

According to our assessment, these states earned high-marks in three key areas: 

1. Developing technical capability
2. Building AI capacity
3. Creating new AI governance structures 

By following this blueprint, other states — and the federal government — can also become leaders.

A Sustainable and Responsible AI Strategy Starts With Strong Technical Infrastructure 

Without modern systems, clean data and appropriate security protocols, even well-intentioned uses of AI can fall flat. Rushing to deploy AI tools without investing in core IT competencies is not only a losing strategy, it can have devastating consequences. 

AI tools have been demonstrated to flag legitimate applicants as fraud, remove eligible recipients from public assistance programs, and embrace harmful biases that disproportionately impact lower-income communities.

States like Utah, for instance, are taking a proactive approach. The state’s Division of Technology Services provides state agencies with access to state-of-the-art platforms and enables secure data-sharing capabilities that feed AI tools with quality data. The division also develops best practices for IT initiatives to ensure AI applications are ethical, legal, and accountable. 

Plus, Utah’s Office of AI Policy — the nation’s first — made it a priority to invest in the necessary infrastructure to support a contained environment for testing and vetting AI models before scaling them across the state.

AI Should Support, Not Replace, State Employees

Tech solutions alone aren’t enough to unlock the transformative potential of AI in government. True AI readiness requires making significant investments in employee training to develop AI literacy, skills, and expertise. Without an AI-educated workforce, state governments won’t reap the full benefits of AI.

That’s why Pennsylvania has aggressively pursued capacity building within its workforce, who form the backbone of government service. The state has partnered with Carnegie Mellon University and InnovateUS to offer AI training programs to state employees. 

Through these partnerships, government workers can access foundational and practical courses on generative AI — making them smarter, faster and better at their jobs. 

Clear AI Governance Structures are Necessary to Ensure Transparency, Accountability and Responsibility

Strong governance leads to clarity around who is in charge, how systems are evaluated, and what corrections are needed when things go wrong. It is what reduces risk and turns AI into a trusted tool for serving the public.

In New Jersey, state leaders established an AI Task Force to study the impacts of AI, think deeply about its applications, and, ultimately, develop a set of recommendations to advance its responsible use. Gov. Phil Murphy also appointed Dr. Beth Simone Noveck as the state’s first chief AI strategist with a mandate to advance responsible AI in the private and public sector.  

These states are charting the path forward — not just to adopt AI, but to do so responsibly. And on that front, our work is just beginning.

Amanda Renteria is the CEO of Code for America, the nation’s leading civic technology nonprofit for more than 15 years.