Universities should embrace ChatGPT

These days, I can’t help but notice how many students on campus are chatting away with ChatGPT.

AI is changing how students learn, usually for the better. It can enhance and augment our learning in ways that don’t resemble cheating. But despite how widely AI is being used by students, university administrators are struggling to find innovative ways to regulate it.

As a graduate student, ChatGPT is like having a personal research assistant. University libraries and online databases are exceptionally difficult to navigate, and at times, ChatGPT provides an elegant way to circumvent them.

Who are the most influential writers on 20th century Russian political economy? This is a question best answered by ChatGPT, unless you fancy spending a few hours combing through bibliographies at the library.

Phrasing is another area where ChatGPT has obvious and non-threatening benefits. What I like about using ChatGPT for phrasing is that students cannot — at least they should not — allow AI to write for them. ChatGPT’s writing is imperfect. At times, its responses are inaccurate and awkward. Handing in an essay written by ChatGPT, then, is typically not in students’ best interest. But it is often helpful to see a sentence written in two or three different ways.

Triangulation is the process of bringing together disparate sources and ideas to produce a novel insight or argument. Here, ChatGPT has been a marvellous asset.

I recently stumbled across two essays from English writers about wandering the streets of London: Charles Dickens’ “Night Walks” and Virginia Woolfs’ “Street Haunting: a London Adventure.” There’s something interesting here but I’m not sure what it is.

Perhaps there’s a longer history of writers who, like Dickens and Woolf, find meaning in a mundane jaunt. Triangulating this theory via a library search could take months. Where would you even begin to look? A search for “English writers” and “walking” is going to come up short.

With ChatGPT, I can find out in seconds if my little theory on writers who wonder has legs.

Here’s the rub: these applications of AI constitute an academic integrity breach at many universities. Yet, this, in my experience, is how most students are already using AI to enhance their school work. And, they’re getting off scot-free.

Many universities have tried to govern AI like mine has, which is to say they forbid any and all use of it, unless a professor says otherwise. Over the summer, McMaster University released similar guidelines, which state, “students should assume the use of generative AI is prohibited unless explicitly outlined by the course instructor.”

This type of policy prohibits the seemingly harmless benefits of AI that make students more efficient learners. It’s also exceedingly difficult, if not impossible to enforce.

Yes, there are tools for detecting AI-generated essays. In fact, it’s not all that difficult for professors to catch an AI-generated essay with only the naked eye. Neither will nab a tactful student using AI in the ways I detail above.

Other universities in Canada have approached the issue of AI with greater nuance. At the University of British Columbia, for example, the school’s AI policy reads, “the use of ChatGPT or other generative AI tools does not automatically equate to academic misconduct.”

Rather, UBC encourages faculty to outline rules governing students’ use of AI early in the term, and to make it an ongoing conversation throughout the year.

In an interview for UBC News, Simon Bates, a Vice-Provost and Associate Vice-President at UBC, says, “banning a tool like this is unlikely to have the desired goal of reducing its use as a way to subvert academic honesty. As with many emerging tools and technologies, generative AI technology comes with both potential benefits and real challenges.”

For now, this is the tightrope university administrators must walk.