Professors adapt to AI in the classroom

Across departments, Middlebury professors are grappling with the implications of artificial intelligence (AI) in higher education. Many professors, including Professor of English Megan Mayhew-Bergman and Lecturer in Chinese Mairead Harris, have modified syllabi to permit AI in specific contexts. 

“I want to make sure we're operating in reality, giving students what they need in this moment, and also pointing our writers and readers to a personal ethos that reflects the world they want to live in,” Mayhew-Berman wrote in an email to the campus. “I like to approach things in a non-sanctimonious way.” 

Bergman plans to host a creative writing chat to promote dialogue about AI usage. She also intends to host analog reading and writing hours. 

Professor of English Timothy Billings has taken a similar approach, distinguishing between “augmentation” and “automation” in the kinds of work that generative AI might do for a student. 

“Education specialists make a distinction between two fundamental uses for generative AI in coursework: 1) augmentation, i.e., uses that “augment” or enhance learning, such as providing explanations of difficult material; and 2) automation, i.e., uses that “automate” or substitute intellectual labor, such as the generation of written assignments or other material for submission. To put it simply, augmentation is allowed; automation is not,” Billings wrote in his syllabus.

While many disciplines are only recently having to develop policies on AI software, language instructors have long reckoned with AI translation tools such as Google Translate. Harris said that, in some ways, the Chinese Department’s AI policy has not changed significantly: at the beginning level, students should not use AI.

In 2023, however, the department reevaluated its policy for advanced students, who are now permitted to use generative technologies for limited purposes.

“Of all of the things that it does, I think it’s pretty good at changing and parsing tone-of-voice. Is this word the vibe that I want? Does this word sound really formal? Does it sound like it’s coming from a specific region?” Harris said. 

Students in Harris’ 300-level class handwrite their initial essays in class. Then, they manually type their notes into a Word or Google document, ensuring that Harris can track changes they make. After drafting, students may use AI tools, provided they submit a log of their interactions with the technology alongside their final draft.

“We also tell them that they can only ask at a specific sentence level,” Harris said. “They can’t take a whole paragraph and copy-paste it into ChatGPT and say, ‘fix the problems.’”

Still, Harris strongly emphasized the importance of generating one’s own language, especially at the early stages of language learning.

“What our goal is for beginning language, is we want students to develop their own Chinese voice in their brain,” Harris said. “[Generative AI] saves a lot of time, and the time that it saves is the whole process that is actually helping students build connections in their own brain that will get them to have an internal Chinese voice. They need time to be wrong.” 

Perspectives on balancing independent learning with AI vary across departments. Professor of Biology Gregory Pask believes that STEM professors may be more concerned with imprecision and environmental impacts of AI.

“I think a lot of people in STEM are looking at the inaccuracies and they might also be more in touch with the environmental impacts,” Pask said in an interview with The Campus.

Pasks noted that when students interpret experiments using AI, the results can be unreliable, and that hearing from the scientists who conducted the experiments is crucial. 

In a follow-up email, Pask added that the writing process is itself essential. 

“It’s designing an experiment, troubleshooting a bug, taking risks in your writing, this is where new synapses form, when the learning happens,” Pask wrote.