The Generative AI in Teaching and Learning Committee (GAITL) report found asymmetric patterns of generative AI use across campus according to a Tuesday Today@Brown message.
According to the report — shared with the community Provost Francis Doyle — 56% of undergraduate respondents and 67% of graduate and medical student respondents reported intentionally using generative AI tools daily or weekly.
The committee noted they found a disconnect between the ways students and faculty use generative AI. They found that while a majority of students who responded to their survey use generative AI to learn new concepts, only a quarter of faculty respondents said they asked their students to use generative AI as a tool for their coursework. When they did so, it was often to have students think critically about AI or use it as a learning assistant.
Doyle charged the committee — co-chaired by Associate Provost for Artificial Intelligence Michael Littman and Assessment and Sheridan Center for Teaching and Learning Transformational Programs Director Eric Kaldor — in March last year with examining generative AI use across the University to identify how it could best support “innovative and equitable teaching and learning,” the report read.
An audit of roughly 3,000 syllabi from academic years 2023-24 and 2024-25 found that more than half did not include a policy on generative AI usage in teaching and learning.
For Nick Burleson ’29, the permissible usage of AI varied across his different professors.
“Some teachers were a little more strict, some weren’t, some of them allowed (AI) slightly, some didn’t,” he said. “Nobody has a clue of what to do and everybody has different opinions.”
Burleson also noted a distinction in the generative AI cases across disciplines. “The thing with humanities is that it needs to be your own thoughts,” Burleson, a computer engineering concentrator, said. “With STEM, the focus is just on building on the next idea to get to the answer.”
The report noted that over 70% of student survey respondents in the life sciences and physical sciences identified as frequent generative AI users, while only 41% of humanities and the arts students identified as frequent users.
Despite differences in AI adoption, students and faculty alike share concerns that over-reliance on generative AI could “reduce long-term critical thinking, have negative cognitive consequences and undermine academic integrity,” Doyle added.
Economics Professor Robert Serrano changed his testing and homework policies based on his experience with students using AI to cheat on a take-home exam.
“It’s clear that take-home exams, at least as we understood them, are a thing of the past,” Serrano said in an interview with The Herald. Serrano also removed the weight of homework from the final grade of his classes to address AI usage on homework assignments.
Respondents have also shared a need “for structured institutional policies and training to develop AI literacy,” Doyle wrote in the announcement.
Based on these findings, the committee outlined a three-phase path forward with near-term, medium-term and long-term recommendations.
For the near-term phase of the roadmap, the committee recommended that the University set baseline rules for generative AI while each department develops “standards for their areas when their expectations diverge from the baseline rules.” The recommendations also call for providing “centralized, enterprise-level GenAI tools” through the Office of Information Technology.
In the next phase, the committee recommends updating the College and Graduate School academic codes to “explicitly address boundaries of AI assistance” as well as providing training to develop AI literacy among University staff.
“AI as a new technology has to be properly understood and properly incorporated into our curriculum,” Serrano said. “The committee, I think, has done a decent job putting together a set of guidelines on all the different aspects surrounding AI in our university.”
The final phase recommends exploring the development of a coalition of peer institutions to set national standards for AI usage in higher education.
This summer, an expanded GAITL Committee will draft potential syllabi templates to discuss with faculty and students in the fall. The committee will also “embark on a listening tour to collect feedback from the community, seeking to ensure that solutions for Brown align with who we are as an academic institution,” the announcement read.
Ivy Huang is a University News and Science & Research editor from New York City. Concentrating in English, she has a passion for literature and American history. Her favorite authors include Marilynne Robinson, Vladimir Nabokov, and Toni Morrison. Outside of writing, she enjoys playing basketball, watching documentaries, and beating her high score on Subway Surfers.
