Generative AI is raising new questions for liberal arts education.

In her Python coding class, Spera allows students to use AI for help, knowing they’ll quickly discover its limitations. “What we knew and the students realized is that AI is not perfect,” she says. “And again, if you don’t have the foundational background in your topic — you not only won’t be able to ask useful questions that can help solve your problem, but you also won’t be able to troubleshoot when it inevitably fails.”

Her teaching philosophy centers on a few key principles: Acknowledge the huge environmental footprint and don’t use AI unnecessarily; don’t let it replace original thoughts; and recognize that AI is biased and imperfect.

“AI is a tool,” she says, “and like all tools, it only works if you know how to use it.” 

Yet he’s not dismissive of AI’s utility. “AI can transcribe my interviews with shocking accuracy, and it helps me tackle administrative datasets that were once too massive and unwieldy to analyze in any reasonable time frame,” he says. “The bottom line is that technology does not replace the human connection at the heart of fieldwork. It amplifies what I can do with the insights I gather there.”

In the classroom — particularly with the role of writing — he says AI poses difficult questions without clear answers, but he offers one prediction.

“The importance of self-regulation in the lives of our students will increase,” he says. “Those who can use AI to enhance their productivity while resisting the temptations to offload the learning process will benefit.”

This creates a novel hurdle. “Navigating this tradeoff is something that no prior generation had to face to the same extent,” he says. “How do you teach this sort of willpower? Another good question to which I have no ready-made answers.”

With respect to research, Driscoll sees practical benefits emerging, particularly as universities and libraries improve their digitization efforts. “Already, I can do so much more primary source research with my students at UR than I could as an undergrad,” she says. Better data sets combined with AI tools that sort and organize information can increase research opportunities, especially for undergraduates.

In her classroom, Driscoll takes a pragmatic approach to how students use AI for writing assignments. She’s developed specific prompts that use AI as what she calls “a second pair of eyes” for student writing. The AI feedback, she concludes, is consistent, helping students identify whether they’ve written a clear thesis statement and providing basic structural guidance.

Driscoll’s careful to emphasize AI’s limitations. She doesn’t allow students to use AI for brainstorming on visual analysis papers — “that’s not the skill that they’re developing.” Instead, the tool helps them learn to edit their own work, mimicking the self-editing skills that experienced writers possess.

Finley-Brook envisions a different path than the one we’re on. “Infrastructure projects for technology like AI should be beneficial to the local host area, and companies need to minimize [environmental] harm — this has not been the trend in data center development,” she says, arguing that a lack of transparency means sustainability commitments require more accountability and independent public reporting.

In her classroom, Finley-Brook focuses on preparing students for the realities of this landscape. “Richmond students can use their critical thinking skills to unpack the claims of tech companies,” she says.

She’s working with fellow CLAAI members on practical resources, including a one-page guide on how to use AI more sustainably. “Not every assignment or project needs to use the largest, biggest, highest parameter model,” she says. “Some projects just need to use a smaller model.”

He integrates AI into his Business and Professional Communication classes using SpiderAI — a UR-built generative AI tool for students and faculty — having students test AI tools toward becoming more critical users. “I want the tool in the loop of human communication practices,” Hocutt says, “not the human in the loop of digital communication practices.”

His advice for everyone is direct: “Immediately start using generative AI because you need to know its strengths and limitations.”

“To the extent that AI turns artists into prompters, I believe it will be catastrophic,” she says. “To the extent that AI can become the basis of enhanced critical practices, exercises, and strange appropriations — that human beings can do something strange and wonderful with it — there is the slightest possibility that it will provide an ecosystem for creativity and reflection.”

In her classes, Bertucci takes a balanced approach. “The goal of my classes is not to be either ‘pro’ or ‘contra’ AI, but rather to raise an awareness about the ideological, technical, and creative problems, potentials, and limitations of AI,” she says.

Her students’ reactions are revealing. “Most students laughed at AI aesthetics — its glossy, flashy sheen — when we analyzed it together in class, as it was so obviously and exaggeratedly kitschy,” Bertucci recalls.

She’s developing what she calls “a commitment to a process over time” as a counter-practice to AI’s promise of immediate mastery.

“An important strand in art production still believes in the necessity of effort, in long years of training and study as part of the formation of a personality and a style,” she says.