I know there’s a lot of concern about the potential issue of using LLMs like ChatGPT and Claude to write essays. And it’s true that there’s a risk they can be used for this. But I worry that we’re placing too much emphasis on a risk that I think largely misses the point, which is that LLMs are so much more than writing assistants.
Students have always been able to use content they haven’t created themselves, and pass it off as their own. Before essay mills there was copy and paste. Before copy and paste there was pen and paper. What’s changed is the scale, speed, and quality of generated content. But in principle, nothing has changed insofar as students have always had the ability to submit work they didn’t create.
What’s new is the ability of an LLM to replace me, and this is the greatest upside of LLMs to support student learning.
If you’d like to have your students start experimenting with language models, but you’re not sure what kind of guidance to give them, suggest that they use language models to replace you.
What exactly do I mean by this? Well, what are the kinds of questions a student would typically ask a lecturer or tutor? And it’s also worth keeping in mind all the reasons they may not ask those questions. Too nervous. Too anxious. Too late at night / early in the morning. Too simple. Too complicated. Too many. Too soon. Too late. And so on.
But they can ask Claude all the questions they would (should) be asking me. At any time. Anywhere.
For example:
- What do you mean by…?
- How do I connect…?
- Can you explain this…?
- What kinds of questions…?
- When do I know…?
- Can you give me feedback on…?
- When is the best time…?
- How do I start…?
- Can you give me another example…?
- Can you help me practice…?
I like to think that I have a certain amount of expertise when it comes to certain subjects, and learning, teaching, and assessment in those subjects. ChatGPT and Claude give students access to a lot of that expertise, and more. And they can access it at any time, any where, and for any thing, without feeling any of the anxiety that comes with putting yourself out there in class (I happen to think that putting yourself out there is one of the best ways to build confidence and grow, but I get why it’s uncomfortable).
I also like to think that I can add value on top of what LLMs provide, so for example, when I know my students well, I can more explicitly link their learning (and my teaching) to their personal experiences. I can also use my own experience to amplify the messages I want to share.
My expertise is also embodied, and comes with the ability to make eye contact, change my tone of voice, shift my body position, and make someone feel heard. LLMs aren’t able to do that (yet).
So when I tell my students to use LLMs to replace me, I do it knowing that I need to up my game in other areas, because access to expertise is now ubiquitous and cheap, and is no longer the limiting factor in someone’s ability to learn.