This story was reported by our media partner VPM News.
Virginia lawmakers want to learn more about how Virginia’s colleges are working — or not working — with artificial intelligence.
Del. Amy Laufer (D–Charlottesville) carried legislation earlier this year to order JLARC, the General Assembly’s research agency, to study AI policies in higher education. As part of the study, JLARC is tasked with developing a model policy for AI use in institutions of higher education across the commonwealth.
“Students are going to want to know what the AI policies are, and students are going to have to use these tools in the workforce,” Laufer told VPM News. “Making sure that we have a competitive edge, I think, is critical, and this is one way to start that discussion.”
The study will focus on academic integrity, data privacy, access, transparency and faculty autonomy.
Andy Pennock, an associate professor of public policy at the University of Virginia who co-chaired a 2023 task force about AI at UVA, said the topics are the right issues to focus on.
In terms of academic integrity, Pennock thinks the key questions shouldn’t just center on cheating, but also whether students are still getting the same value out of the university experience that they got before AI.
A key line in the 2023 report from the AI task force, Pennock noted, is that “ubiquitous access to generative AI has the potential to allow students to learn far more than they did last year, but also means that they could learn far less.”
In terms of access, Pennock thinks about who does and doesn’t have access to different AI tools — including more advanced models like Microsoft’s Copilot Premium, for example — within the UVA community.
The university has been making some AI tools available to faculty and students, and the school is in the process of expanding access to additional tools. Pennock’s not sure that would be financially feasible for other public universities in the state.
“I imagine that will be even more complicated at the less well-resourced universities, particularly the community colleges and some of the ones that receive less state funding,” Pennock said.
Members of the UVA community have been thinking about AI for years. The 2023 AI task force report led to the development of some general guidance about AI usage by students and faculty.
The guidelines don’t bar AI use university-wide, but leave the decisions up to instructors, who are told to make explicit in their syllabi and assignment descriptions which uses of AI (if any) are permitted in coursework and which are prohibited. If students do use the technology, they are told to cite its use in their work.
Leo Lo, dean of libraries at UVA and special advisor to the provost on AI literacy, told VPM News, “There are efforts to send a more university-wide AI vision. But for the last several years, different schools have been experimenting, trying different things.”
There are also multiple ongoing university-wide committees at UVA dedicated to thinking about AI use in learning, research and policy design. Lo said there are AI-related initiatives taking place all over the university — including the library’s AI Literacy and Action Lab, which he co-founded.
“At this point, there’s not enough evidence to say, ‘This way would help you or hurt your learning,’” Lo said. “So, we need more of these experiments, assignments, different ways of using AI, and then we can learn from that.”
Natasha Heny, a professor in UVA’s school of education who co-chaired the AI task force at UVA in 2023 along with Pennock, told VPM News in a recent interview: “Being informed is the most important thing about the way we’re thinking about AI: making informed decisions about AI.”
She hopes the JLARC study will consult experts not just in AI, but also across various academic disciplines.
“I want the study to be robust, and I also want it to be able to adapt,” she said. “It’s changing so quickly, and we need a structure in place to continue to study and to continue to learn.”
Ryan Wright, special advisor to the provost on AI, said some of the ongoing discussions can continue through the Commonwealth Artificial Intelligence Institute, which is in its planning and development stage. UVA is leading the statewide initiative, which has the stated goal of “positioning the state as a national leader in responsible and economically-transformative deployment of AI.”
UVA received a $500,000 planning grant to get the Institute off the ground, and to “coordinate with other universities and stakeholders to develop a statewide ecosystem in AI.”
“This conversation isn’t about UVA,” Wright said. “This conversation doesn’t happen in a single university. This conversation needs to happen across the commonwealth with lots of different stakeholders.”
He added: “The science hasn’t been settled around this stuff. The best practices haven’t been settled, so we need to come together and we need a long-term look.”
George Mason University has also been thinking a lot about AI, and has developed guidelines for faculty, students and researchers similar to UVA’s. The Northern Virginia university — which hosted a statewide AI summit last year and put together a reference guide based on key takeaways — is also involved in the CAII.
Both universities have been focusing heavily on training faculty to ensure that if they’re going to use AI, they know how to use it responsibly and ethically.
“We are at the beginning of a change of culture,” said Amarda Shehu, GMU’s first chief AI officer. “We just have to be a little bit better in terms of taking that plunge and experimenting. How should the classroom change now in the presence of this technology?”
Padmanabhan Seshaiyer, a mathematics professor and co-author of the GMU AI reference guide, said higher education needs to rethink what authentic learning looks like.
“The goal should be to create that environment where students can demonstrate what they themselves understand, what AI has contributed, and how they verified the result,” Seshaiyer said.
The JLARC study hasn’t yet kicked off, but staff say they plan to consult experts in Virginia and nationally. The agency is also tasked with recommending AI tools, curricula and other resources that could be included in a “statewide clearinghouse” for educators and students at institutions of higher education, as well as for the public at large.
Kimberly Sarte, an associate director at JLARC, said staff will be able to apply some information gathered as part of a forthcoming study about the use of AI in state government towards the higher education study, particularly as it pertains to data privacy.
“Some of these concepts are just general practices that anybody using AI needs to be aware of and address,” she said.
Meanwhile, other areas are more specific to higher education, like academic integrity and faculty autonomy.
Justin Brown, also an associate director with JLARC, said that because the situation with AI is very fluid right now, the commission will focus less on status information and more on “enduring and structural principles that need to be carefully thought through and understood for the long term if you’re going to have smart governance about this.”
The report is due by early 2028.