Strengthening the state and local workforce with AI literacy

State and local agencies are navigating an increasingly complex AI environment. Federal guidance, grant conditions, state executive orders, local procurement rules and agency-specific policies all shape how AI is evaluated, adopted and governed, often by teams already managing heavy workloads and limited resources.

For Tony Holmes, practice lead for solutions consulting at Pluralsight, the answer is not to stand up a separate compliance effort for every new requirement. Instead, agencies should focus on the throughline many frameworks have in common: applying AI to real tasks, evaluating outputs and using the technology responsibly.

“Build the literacy layer once, and every framework’s requirements get structurally easier to meet,” he said. “Then you've built the thing they actually are asking for underneath the paperwork.”

This idea is supported by the Department of Labor’s AI literacy framework, which Holmes noted was created with state workforce agencies, local workforce boards, labor commissioners and community colleges in mind. While part of the framework calls for a basic understanding of how the technology works, much of it centers on bolstering human judgment.

“It’s voluntary guidance addressed to state and local governance, and it already starts from the point of treating literacy as judgment and building for agility,” Holmes said. “It wasn’t written at state and local government, it was written to them. Federal agencies are actually borrowing SLG’s playbook.”

Discernment is especially important for state and local agencies as they use AI to support critical resident services, including permitting, benefits administration, workforce programs, transportation and public safety. In each of those areas, employees need to know the appropriate time to use the technology, when to question it and when human review is required.

“A framework document can’t catch a confidently hallucinated citation in a draft decision memo, or notice an agent that’s gone subtly wrong. It can’t apply context-specific caution in a high-stakes case, but your workforce can if you give them the literacy to evaluate that output,” Holmes said. “Literacy is the governance, and it makes it operationalized.”

Operationalizing AI literacy through Pluralsight AI Academy

That’s where Pluralsight AI Academy comes in. The program is designed to help agencies embed AI literacy across operations through an end-to-end learning journey. Coincidentally, AI Academy’s architecture was finalized at the same time the Department of Labor published its literacy framework, and the two approaches closely mirror each other (“Great minds think alike,” said Holmes). Both focus on experience over theory and emphasize practical, judgment-centered learning that can evolve as AI tools and use cases continue to change.

“AI Academy is a maturity journey,” Holmes said. The program is tailored to each organization and individual, beginning with assessments to establish a readiness baseline, then moving through three levels.

• AI literacy. All employees are provided a shared language around AI’s capabilities, limits, risks and responsible use. The goal is not to make every employee an AI expert, but to create a common floor of understanding across the organization so teams can discuss AI-enabled opportunities with the same vocabulary.

“That floor is how pilot success translates, because now non-technical teams or people not part of the pilot cohort are able to identify the places that artificial intelligence can help them,” said Holmes.

• AI productivity. At this stage, agencies move from awareness to application, using hands-on labs, role-aware learning and contextualized seminars tied to the tools and workflows the agency actually runs.

“This is where the sorting happens,” Holmes said, “because it helps people understand how to automate the toil, and how to augment the friction where you still need a human in the loop.”

• Agentic AI. Smaller cohorts learn to prototype agents against real agency contexts, paired with safety, testing, readiness and governance considerations.

“The ordering matters because agents before literacy means outputs that nobody can tell are correct or useful,” Holmes said, noting that more autonomous tools will require more human expertise and sound judgment.

The program also offers a variety of learning pathways that serve different purposes.

• On-demand learning builds vocabulary and mental models.
• Labs build instincts through practice, trial and error.
• Seminars connect lessons to the tools and workflows agencies actually use.
• Workshops help produce working artifacts, not just completion certificates or hours watched.

“AI can obsolete standard courses faster than you can certify people through them,” Holmes said. “The curriculum has to be built around the one thing that doesn't move, which is workers’ judgment about their own work.”

The state and local advantage

AI education and adoption can be daunting, but state and local agencies already have a unique advantage. 

While many AI pilots in larger agencies show promise, initiatives often stall when they move from testing groups to the rest of an organization’s employees, who may not understand where or how those capabilities fit into their workflows. Holmes explained that the smaller size of state and local governments allows them to be more agile.

“They’re in a significantly better position for AI,” he said. For example, a federal department of 40,000 may need long pilot cohorts and phased rollouts to close the gap between trained and untrained employees. 

By contrast, “a 2,000-person state agency or large county can easily reach saturation across the whole cohort, not a sample,” Holmes said. “The saturation is what converts those pilots into execution, because there’s no untrained population left to have misunderstandings or friction across.”

Pluralsight AI Academy can help expedite and streamline implementation by giving agencies a structured way to build judgment and applied skills in the context of their teams’ real workflows.

“State and local don’t need a new strategy,” Holmes said. “They need to refocus the conversation toward literacy and workforce-wide understanding. The good news is that they’re agile, and what artificial intelligence applications require now is absolute agility, so it's a conversation they can start having this quarter, this year.” 

Learn more about how Pluralsight AI Academy can help your workforce prepare for AI.