Hello and welcome to Best Of Maclean’s.

My students are using ChatGPT to write papers—and I support it

Boris Steipe is a professor in the University of Toronto’s biochemistry department, where he witnessed ChatGPT go from a curiosity to a phenomenon among students in a matter of weeks

I’ve been a professor in the University of Toronto’s biochemistry department since 2001. Last fall, I taught a first-year course on the foundations of computational biology, a field that applies computer-science principles to the study of biological processes. For the final assignment, I asked each of my 16 students to examine data on some genes involved in damage repair in human cells and write a short report on their findings. There was something off about two or three of the responses I received. They read like they’d been written by students who were sleep-deprived: a mix of credible English prose and non-sequiturs that missed the point of what I was asking. A few weeks earlier, the AI chatbot ChatGPT became available to the public, but it was so new that it never occurred to me that a student could be using it to help them with an assignment. I marked the reports and moved on. 

In the following weeks, I saw ChatGPT change from a curiosity into a phenomenon. It came up in every discussion I had with fellow faculty members, across all disciplines—from the humanities to data sciences. My colleagues wondered how one could tell whether a student used the AI to answer questions, and many were concerned with how it might enable plagiarism. What if a professor suspected a student had used ChatGPT but couldn’t prove it? A plagiarism allegation is no small thing: you can’t risk ruining a student’s academic career on a hunch. I thought again about that December assignment and realized those off-putting answers might have been my first brush with ChatGPT. 

I directed the university’s bioinformatics and computational biology program for 15 years, and often wrote programs for my own lab, so I had been learning to “talk” to computers for years. I signed up for a ChatGPT account and became intensely interested in finding solutions to this new problem. I figured it made no sense to ban ChatGPT within the university; it was already being used by 100 million people. Over the Christmas break, it hit me: as professors, we shouldn’t be focusing our energy on punishing students who use ChatGPT, but instead reconfiguring our lesson plans to work on critical-thinking skills that can’t be outsourced to an AI. The ball is in our court: if an algorithm can pass our tests, what value are we providing?

Continue Reading