Professors Explore Constructive Use of A.I. in Classrooms
Across universities, the arrival of generative artificial intelligence (A.I.) in classrooms is sparking both concern and innovation. While some educators remain staunchly opposed to the use of tools like ChatGPT, others are experimenting with ways to incorporate A.I. into teaching methods that enhance rather than replace student learning.
At Barnard College in Manhattan, Professor Benjamin Breyer is at the forefront of this educational shift. Despite the college’s general restriction on the use of generative A.I. in its first-year writing program, Breyer has received special permission to test its integration in his classes. His goal is to explore how A.I. might support students in developing academic writing skills.
Balancing Innovation and Caution
The writing program at Barnard, under the leadership of Wendy Schor-Haim, maintains a cautious stance. Schor-Haim leads screen-free classes and relies on traditional annotation techniques to teach critical reading and writing. She has never used ChatGPT and believes A.I. performs poorly in the very skills writing courses aim to teach. However, she has allowed Breyer to experiment within his classroom, signaling a willingness to examine alternative approaches.
“Students tend to use A.I. to circumvent learning rather than enhance it,” Schor-Haim said. “And it is very, very bad at that work.”
A New Kind of Classroom Assistant
Breyer, a veteran professor at Barnard and Columbia University, has worked alongside former college software developer Marko Krkeljas to create an A.I. chatbot named Althea. Designed to function as a supplemental tutor, Althea assists students in developing thesis statements and practicing annotation. Unlike commercial bots, Althea refuses to complete writing tasks for students. Instead, it provides prompts to refine their thinking and writing.
“A.I. may help with expression, but the core idea — the hardest part to teach — remains within our domain,” Breyer explained. “This isn’t a threat. It’s a tool.”
To ensure Althea aligned with his own teaching methods, Breyer trained the bot using transcripts of his lectures, sample student work, and course materials. He also tweaked its tone to avoid being overly complimentary, aiming for constructive, honest feedback.
Mixed Results and Ongoing Refinement
Breyer’s initial findings were mixed. After three semesters of comparing writing outcomes between students who used Althea and those who didn’t, the results showed that non-A.I. users performed slightly better. Rather than abandoning the project, however, Breyer improved Althea’s focus, training it over the summer to ask more targeted questions and avoid disrupting students’ thought processes.
The changes paid off. In the most recent semester, students using Althea outperformed their peers on writing exercises. More importantly, Breyer noted that the process of refining the chatbot enhanced his own teaching by helping him design better questions and reflect on his instructional approach.
Other Educators Join the Experiment
Breyer isn’t alone in his efforts. At George Washington University, Professor Alexa Alice Joubin has also developed a custom A.I. assistant to support her English classes. Her bot, created in collaboration with a computer science student, helps students summarize readings and formulate research questions. The bot responds in bullet points and is designed to protect privacy and intellectual property, unlike many commercial alternatives.
“This is our sandbox to see what A.I. can and cannot do,” Joubin said. Her tool is open-source and available for professors around the world to adapt for their own classrooms.
However, challenges remain. Joubin observed that some students use the bot to prepare for class at the last minute, indicating they may be skipping the required readings. This highlights the ongoing tension between using A.I. as a learning tool and relying on it as a shortcut.
Institutional Shifts and Broader Adoption
At Columbia University, history professor and Vice Dean for A.I. Initiatives Matthew Connelly encourages students to use A.I. for research but strongly discourages its use in writing. “Writing is thinking,” Connelly said. “And A.I. can be pernicious in that process.”
He is currently leading a study to understand A.I. use among first-year writing students and faculty. Connelly believes that, while this is a time of uncertainty, it also presents an opportunity to reimagine how education is delivered. “It’s a very scary time,” he said. “But I actually think this is an incredibly exciting time in higher education.”
Legacy academic organizations are also adapting. A 2024 working paper from the Modern Language Association’s task force recommended that first-year writing courses teach students how to engage with A.I. critically and responsibly.
Looking Ahead
As technology continues to evolve, the presence of A.I. in the classroom seems inevitable. Professors like Breyer, Joubin, and Connelly advocate for proactive engagement to ensure that these tools are used effectively and ethically. Rather than resisting change, they are shaping how the next generation learns to think, write, and communicate in an A.I.-enhanced world.
This article is inspired by content from Original Source. It has been rephrased for originality. Images are credited to the original source.