Virginia’s new AI pilot program to cut regulations and raises questions on efficiency, transparency

RICHMOND, Va. (WRIC) -- The buzz around artificial intelligence (AI) isn’t just in technology circles anymore — it’s now making its way into Virginia’s state government.

Gov. Glenn Youngkin recently signed an executive order in July launching a pilot program that uses AI to review and shorten state regulations, aiming to make government more efficient.

But experts are cautioning Virginians that the move isn’t as straightforward as it sounds.

Max Stephenson, Jr., director of Virginia Tech’s Institute for Policy and Governance and a professor of public and international affairs, spoke with 8News about the challenges and risks of applying AI to regulatory reform.

“This initiative is framed as making government more efficient by making these regulations shorter,” Stephenson said. “But regulation isn’t just paperwork to trim. It’s the exercise of administrative discretion, balancing multiple values like safety, fairness, and effectiveness.”

Stephenson highlighted how complex these regulations are and cited how often they require nuanced language to address diverse real-world situations.

“One can shorten things and make them more concise and clear, but you can also lose nuance, context, and different dimensions of a question,” he said.

“He [Youngkin] has not released a methodology to my understanding of how this was done or coded, or why it was coded in the way it was coded," Stephenson said, questioning the methodology behind the AI program. "Or why, relatedly, shorter is necessarily better.”

PREVIOUS :Youngkin launches first-in-the-nation AI pilot to streamline Virginia regulations

In a statement about the executive order, Youngkin said this will benefit the Commonwealth, specifically its economy.

"We have made tremendous strides towards streamlining regulations and the regulatory process in the Commonwealth. Using emergent artificial intelligence tools, we will push this effort further in order to continue our mission of unleashing Virginia’s economy in a way that benefits all of its citizens." Gov. Glenn Youngkin

Stephenson emphasized the challenge of the new program and compared regulations to health inspections at restaurants.

“An inspector has to balance efficiency versus effectiveness, and in some cases equity — making sure employees aren’t unsafe in the name of efficiency,” he said. “That’s administrative discretion. AI isn’t ready to replicate that.”

While cautious, Stephenson believes AI could have a place in government — if applied carefully.

“I think of this as a very powerful tool, but certainly in regulation, we still need human judgment,” he said. “You can program algorithms for routinized tasks, things that are uniform and repetitive. But for complex discretionary choices, we’re not there yet—and maybe never will be.”

Stephenson said the importance of transparency and accountability is crucial.

“So far, we don’t know the methodology," he said. "That would have to be released so people can make reasoned judgments about what it actually does. And there needs to be accountability—someone who is willing to be held responsible for the results and explain the judgment criteria used.”

Stephenson also emphasized who should be involved in these conversations.

“We need a broad conversation — not just on the executive side, but in the General Assembly, in committees responsible for lawmaking and oversight," he said. "The judiciary will also get involved as people challenge AI-based decisions in court.”

The stakes are especially high for vulnerable populations.

“Regulation often protects the public from harmful actions,” Stephenson said. “For example, the ADA mandates curb cuts, which cost money but ensure accessibility for people with disabilities. If we evaluate everything only on efficiency, these important protections can be lost.”

On the issue of public trust, Stephenson expressed skepticism about whether the AI initiative will improve confidence in government.

“This initiative builds on skepticism that government is inefficient — whether that’s true or not. It plays to a political trope," he said. "It isn’t about building trust or effective governance, as far as I can tell. It may actually make trust worse.”

However, he sees a potential role for AI in policymaking dialogue.

“AI could help us frame policy questions, consider scenarios, and develop nuanced analysis," he said. "But that’s very different from what this initiative is doing.”

“We need to know exactly what was done — what regulations, which agencies, what methodology. Then we can ask: What assumptions were made?" Stephenson said. "What was lost or gained in making regulations shorter? Is it more efficient, more effective, more equitable? And can we even know that based on how this was done?”

“People should be able to ask these questions about their local dry cleaner, restaurant inspections, or highway safety regulations — and deserve clear answers,” he added.

As the state experiments with AI in government, Stephenson’s cautionary words remind us that governing isn’t just about cutting red tape, but balancing complex, sometimes competing needs to serve all citizens fairly.