If you read any education publications, you’ve probably heard about ChatGPT, whether you want to or not. Writers’ reactions range from excitement to doomsaying to dismissal, but everyone seems to have something to say.
A common topic of discussion is how ChatGPT will impact student research and writing. Although I haven’t encountered this myself, I’ve heard stories of librarians being approached by students with citations they want to track down, only to learn that the student generated the citation using ChatGPT and the article does not actually exist. I’ve also heard about schools considering banning the software. Nothing of the sort has happened at my own institution – in fact, there have been no statements, mandates, or guidance put forth at all. If people at my college are discussing ChatGPT, it is only behind closed doors.
Because of this, I was intrigued to come across a recording for a LiLi Show and Tell webinar from February 2023: Using ChatGPT to Engage in Library Instruction? Challenges and Opportunities by Ray Pun. In the presentation (embedded below) Pun introduces librarians to some basics about how the tool works, discusses benefits and risks associated with its use (with a focus on BIPOC and international students), explores how ChatGPT complicates citation and attribution, and offers ideas for activities integrating ChatGPT into information literacy instruction, with an eye toward helping students approach the tool more critically.
Pun takes pains to clarify that the content of the presentation will become outdated quickly (in fact, GPT-4 was released after the recording), but I still found a lot of the information thought provoking. In particular, his point that suspicions over ChatGPT use could fall more heavily on BIPOC and international students gave me pause. Instructors relying on stereotypes might have lower expectations for, say, non-native English speakers’ writing, and if the writing exceeded those expectations, they could falsely accuse the students of using AI assistance. It’s a reminder that educators need to remain aware of their own biases.
Although I don’t have a ChatGPT account and am not currently planning on making one, I was also taken by one of Pun’s suggested activities: having the AI generate a reading list of sources for a topic and then asking students to critique the list. Are all the sources real? Do the selections make sense? What’s left out? It’s a good way to show students the limitations of the AI and get them thinking about the broader limitations of ‘canonical’ research – the AI’s selections will likely recreate hegemonic biases in whose research gets the most attention.
Whether we like it or not, and whether commentators’ positive and negative predictions end up being overblown, AI is something we as librarians will have to deal with for the foreseeable future. Presentations like Pun’s are a great way for us to start thinking about how to handle it.