Middlebury does not currently have an institutional policy dictating how artificial intelligence (A.I.) should be used in the classroom, which leaves the decision up to individual professors. In an attempt to encourage intentional A.I. use in academic instruction, the Center for Teaching, Learning, and Research (CTLR) and the Office of Digital Learning & Inquiry (DLINQ) recently hosted a series of workshops for professors and peer writing tutors that focused on strategies and techniques for implementing A.I. as a tool for learning.
Author and educator John Warner also gave a talk on campus on March 11, entitled “Resist, Renew, Explore: How to Think About Human Writing and Learning in the Age of A.I.”
The Writing Center, located in the CTLR, organized a workshop with Associate Professor of Writing and Rhetoric Héctor Vila on Feb. 26 for peer writing tutors on the use of A.I. The event focused on the benefits of tutors using A.I. to identify the issues with a student’s writing, allowing them to work closely with students to make actual changes to their papers during sessions, instead of the students making those edits later.
However, beliefs on the extent to which A.I. should be used in the classroom vary. Cole Siefer ’25 and Aren Lau ’27, two peer writing tutors that attended the workshop, expressed some of their concerns afterwards, believing the tutoring process is more impactful without artificial intelligence.
“If students don't need to do their readings, or don’t need to understand the readings themselves, what's the significance of writing about them?” Siefer said. “I guess that’s a pretty philosophical question, but I think it's worth considering, if students don’t need to do the leg work to understand the material that they're learning, what value is their writing?”
Lau echoed Siefer’s hesitations about the emerging technology’s use in the classroom.
“Of course A.I. can help you understand and digest things, but isn't the process also part of it?” Lau said. “It's not just about taking in the content. It's about engaging with the content. You can't totally understand something unless you actually engage with it and A.I. puts a barrier in between that engagement.”
The workshop also brought to light the challenge for writing tutors to implement AI tools while adhering to the Honor Code, which does not assert a uniform A.I. policy. The Middlebury Handbook currently permits the use of artificial intelligence such as ChatGPT when permitted by an instructor, but provides no guidance to the instructors themselves.
“Any use of generative A.I. tools that is not expressly permitted or approved by an instructor may be considered cheating and/or plagiarism,” the handbook states.
Siefer stressed that the lack of clear direction in instructor policies places writing tutors in a difficult position.
“One of my biggest concerns was going against the Honor Code in my writing tutoring sessions. I don't want to have to read a student’s syllabus when I’m helping them to know if I can use [A.I.] or not,” Siefer said.
Despite student and faculty concerns, Vila expressed his belief in the widespread importance of AI and the necessity of adapting ways of teaching and learning to take advantage of it.
“A.I. itself is like a bowling ball that hits all the pins. So it's hitting every department, every structure, and if you are a static structure, you're going to get blown up,” Vila told The Campus.
Vila frequently incorporates tools like NotebookLM and ChatGPT into his own teaching. NotebookLM and ChatGPT are both examples of generative A.I. (as opposed to traditional A.I. like recommendation assistants on Netflix and Amazon), which means they can create completely new content based on patterns they learn to pick out from existing data. His students use NotebookLM to peer edit each other’s essays by asking the A.I. tools to generate a summary of their partner’s argument, and if the A.I. summary differs from what the writer intended, the student can then rewrite and clarify the point on their own. Vila explained that NotebookLM is especially useful for teaching and learning because it does not rewrite content the way that ChatGPT can.
“It's strictly a tool to make you think deeper and get more engaged in the minutia of your work,” Vila said.
On the issue of academic integrity, Warner’s March 11 talk emphasized the need to reevaluate our education system as a whole, arguing that education has become transactional, where “students are turned into customers and generative AI satisfies that transaction.” Warner suggested that in order to discourage students from abusing AI, it is critical to help them see the value of education and embrace curiosity.
As artificial intelligence continues to evolve, its future in Middlebury classrooms remains uncertain. While it possesses the ability to make learning more efficient, it also can take away the critical thinking that is at the heart of intellectual curiosity and learning.
