Closing the AI Skills Gap in Government: A Practical Path for State and Local Workforce

Artificial Intelligence (AI) is rapidly reshaping how government agencies manage data, deliver services, and respond to the needs of the community. Yet for many state and local governments, the technology is advancing faster than the workforce’s ability to confidently and responsibly use it. 

Core Challenge: AI Adoption Tuned to State and Local Missions and Constraints

AI modernization is frequently pursued through large-scale infrastructure investments and enterprise-wide process improvements, supported by multiyear funding, centralized IT organizations, and long-term modernization programs. At the state and local level, AI adoption follows a different operational model, shaped by leaner teams, more variable funding, and a focus on maintaining essential services, and trying to enhance the lives of the dynamic and unique communities they directly serve.

These differences affect how the workforce adopts AI and how training is delivered. Federal agencies may develop AI strategies aligned to enterprise policy frameworks, while the more varied nature of state and local adoption means city planners may use AI to analyze zoning forecasts, HR functions including performance reviews, and state transportation departments will lean towards optimizing predictive maintenance. The result is a set of dynamic and complex cross-agency training needs that must be addressed with extremely limited staffing and resources. 

These challenges are reflected in recent public-sector research showing lower reported AI usage among state and local workers, with 51% indicating daily use compared to 64% of federal workers. This gap reflects differences in both resource availability and policy structure.

Agencies increasingly operate within centralized AI governance models, guided by executive orders and enterprise-level strategy goals. State and local governments, by contrast, must develop governance approaches independently, resulting in wide variations based on local priorities, capacity, and constraints. 

Practical Training Approaches That Fit State and Local Realities

Given the structural constraints facing state and local governments, workforce development efforts must be practical, low-barrier, and closely tied to day-to-day responsibilities. Three priorities, in particular, can have an outsized impact in boosting AI adoption:

• Ground AI Training in Real Workflows: State and local staff are more likely to engage with AI when they see clear relevance to their roles, whether that means improving permitting and licensing turnaround times, accelerating eligibility reviews in human services, or helping finance teams identify anomalies across large datasets. Showcase the value of finding the needles in haystacks that could never be found by even a team of human operators. Applied, role-specific training builds confidence and trust more effectively than generalized instruction on AI capabilities.
• Leverage Existing Learning Structures: Most agencies already operate learning management systems, internal professional development programs, and informal peer-to-peer knowledge networks. Adapting these existing mechanisms to include AI literacy allows agencies to scale learning without incurring significant new costs. Peer-led learning, in particular, is powerful in helping employees learn from AI-savvy colleagues who understand local operational realities and constraints. 
• Emphasize Practical Judgment Over Technical Depth: AI literacy does not require employees to become technical specialists or to understand how models are built. Instead, training should focus on practical understanding of where AI is well suited, where it introduces risk, and when human judgment must remain central. This clarity is especially important in state and local environments, where stringent AI governance frameworks may still be evolving and employees require reassurance that technology is being introduced to support their work, not to replace it. Key themes to focus on include:
  • Goals: What is the problem that I am using AI to solve?
  • Purpose: How much autonomy am I providing it (Toy, Tool, Teammate, Autopilot)?
  • Efficient: What is my AI good at (pictures, writing, audits, recommendations, etc.)?
  • Accuracy: Where is the data sourced and who can manipulate it?
  • Relevant: Is my model current, i.e., how often is the model trained and RETRAINED?
  • Safety: What guardrails or safeguards are in place (Privacy-, PII/HIPAA-compliant)?

Taken together, these priorities allow state and local governments to build AI readiness responsibly in ways that align with both mission demands and resource realities, without overwhelming the workforce or budgets.

Embedding Governance and Responsible Use Into Training

Responsible AI use must be well-integrated into training protocols. Without governance safeguards, even well-intentioned AI applications can produce unintended consequences. For example, translating public meetings into multiple languages with AI can expand access, but without proper validation protocols, it can also introduce errors that confuse citizens or misrepresent official content. 

These risks underscore the need for AI training to include clear guidance on data usage, risk mitigation, and when not to use AI without appropriate oversight. Cross-agency collaboration is key; sharing lessons learned on governance approaches, training materials, data safeguards, and vendor engagement can reduce duplicative effort and accelerate learning.

AI policies and applications continue to evolve, requiring training programs to do the same. Effective workforce development depends on continuous learning environments that track how changes in technology and policy affect the way employees use AI in their daily workflows. These feedback loops are more effective than static curricula, which can quickly become outdated as tools and requirements change. 

For state and local governments, AI adoption in the workforce succeeds when training programs focus on the reality of broader missions and tighter resources that agencies face. This calls for training models that are rooted in everyday workflows, local missions, and existing learning structures that emphasize peer-to-peer learning. Supported by strong governance and ongoing curriculum optimization, these priorities help ensure that AI adoption effectively teaches the workforce and improves state and local public service delivery.

Jason Dunn-Potter is a Solutions Architect for the US Public Sector at Intel Corporation. He works directly with Government agencies and Industry partners to solve complex problems by providing trusted thought leadership and candid feedback.

Photo by Luke Jones on Unsplash