By Courtney Mitchell. Photos by Melanie Busbee.
Generative artificial intelligence, or GenAI, seems to be everywhere in 2025. Company websites offer chatbots that answer questions and retrieve information. Google returns an “AI overview” at the top of searches. Software like OpenAI’s ChatGPT gives conversational answers to questions, writes emails, summarizes complex documents and can even tell jokes. Social media is awash in AI-generated art and music, and GenAI skills are a requirement for more and more jobs.
GenAI is a type of artificial intelligence that draws upon huge sets of existing text, data and other information. When a human issues a prompt, the GenAI software uses these models to craft new text, images, computer code and more. GenAI reached the mainstream with the release of ChatGPT in 2022, evoking both curiosity and caution.
“As another technology in a long, historical line that helps us make sense of the world and create new knowledge, it fell squarely in the Libraries’ portfolio.”
— María R. Estorino
As Vice Provost for University Libraries and University Librarian María R. Estorino listened to her campus peers—school and unit deans and members of the administration—discuss what GenAI might mean at Carolina, she knew the Library was well positioned to meet the moment. The University’s information experts who could help campus navigate the latest in AI were already right here.
“GenAI is something that’s going to touch all of us,” she says. “I wanted us to be part of that conversation from the beginning and not just wait to see what happens.”
Estorino recalled scholar Alison Gopnik’s formulation of AI in a 2022 Wall Street Journal column as a “cultural technology, like writing, print, libraries, internet search engines or even language itself.” Just as a student in the 1990s came to the library to learn about the internet, today’s students should know they can come to the library to learn about AI.
“As another technology in a long, historical line that helps us make sense of the world and create new knowledge, it falls squarely in the Libraries’ portfolio,” observes Estorino. “There’s a larger information literacy framework that grounds our work. We want people to understand not just how to find information, but how to evaluate it and how to use it.”
An AI roadmap
There are big questions librarians can be part of answering: How can students use GenAI ethically without plagiarizing? How can humans parse out the inherent bias of information that comes from a large language model? Can GenAI save time on operational tasks, like synthesizing information? How can GenAI be used responsibly?
As part of the Libraries’ strategic framework and priorities, a small team came together last summer and developed a roadmap to guide the organization’s AI efforts. Leading the way are Michelle Cawley, associate University librarian for health sciences and director of the Health Sciences Library; Tim Shearer, associate University librarian for digital strategies and information technology; Amanda Henley, head of Digital Research Services; and Cole Hargrove, Betty Debnam Hunt and Richard M. Hunt Technology and Discovery Fellow.
The group has focused on preparing all parts of the University Libraries for engaging with AI. They have created learning and brainstorming opportunities for Libraries staff, steadily building a community of AI-savvy information professionals. They are tracking and bolstering AI-driven projects around the organization and laying the foundations for a future that empowers students and faculty to use these technologies.
“I think it really behooves us to get on top of this and understand exactly how it’s going to impact research—how researchers can harness GenAI and how we can improve our own operations in the Libraries,” says Henley.
In many ways, the Libraries had a head start. Henley’s unit began providing data visualization, programming and data analysis services about a decade ago through the Research Hub, now Library Data Services. Most recently, her team used GenAI to help art professor Kathryn Desplanque compose descriptive text for some 500 political cartoons from 18th- and 19th-century France.
At the Health Sciences Library, which supports all five health affairs schools and the UNC Medical Center, librarians are investigating how GenAI applications can speed the time-intensive process of conducting a systematic review. And last year, librarians from across the Libraries began teaching GenAI workshops; the popular sessions have attracted more than 700 participants so far.
What comes next?
Generative AI is still in its early days and is evolving quickly. As the Carolina community engages with the Libraries’ programs and grapples with this cultural technology, the Libraries plans to stay nimble and responsive.
It’s also important to leave room for doubts, concerns and criticisms, says Estorino. Data centers that host the cloud-based, large-language models that fuel GenAI gobble electricity and water to cool servers. Estorino says there have been discussions in the field around ways to lessen the environmental impacts. There’s so much no one knows, yet—but the Libraries will be alongside campus partners every step of the way as they figure it out together.
“Right now, we’re thinking a lot about removing barriers for people who want to engage,” says Estorino. “As our services and generative AI both mature, we’ll have more clarity about using these tools. We want to be the obvious campus partner for researchers, students and instructors—anyone who wants to work with GenAI or any data-intensive service.”