Checking In with AI at College

How goes the revolution?

Just this week, Inside Higher Ed has already published four articles about the implications of generative artificial intelligence for my field. There’s the usual array of techno-enthusiasm — AI can democratize admissions counseling and improve productivity — but beneath it all is lingering uncertainty about how to use such a rapidly-developing tool and whether it’s a good idea in the first place.

Good news on that front, however: the American people trust colleges and universities far more than social media and politicians to use AI responsibly!

If only I could be so sure. Like IHE contributor John Warner, I think there are some obviously inappropriate applications of generative AI, uses that would violate values that should be universal in higher ed. But even as I try to hold those lines, my own limited experience in teaching students with access to these tools makes me realize just how many shades of gray are popping up.

Briefly, here’s a survey of how I’ve seen generative AI being used since it burst on the scene.

Non-Use

What’s struck me most from the past academic year is that, in general, my students write (and write well) without any obvious use of AI. Now, it’s entirely possible that they are using it in subtle ways that I’ll get to later; it’s just not raising any red flags yet.

But I’m actually impressed — and relieved — to realize that students are not only perfectly capable of writing well at this level, but that most of them take at least somewhat seriously the value of working at it. At some level, they get that writing is not only a useful skill honed through practice, but that it gives us a chance to explore ideas, engage in conversation, clarify our beliefs, and express ourselves creatively. In that sense, writing is ultimately autobiographical, something that no AI can do for any human.

This once, I thought it would be appropriate to use AI to generate images for the post. This is what deepai.org created from the prompt “frustrated writing professor.”

For the most part, I think my instinctive response to teaching writing in an AI age was right on. As much as possible, I’ve tried only to assign writing prompts that depend on students integrating specific sources that demand close reading and/or demonstrating personal engagement and reflection. And I’ve increasingly tilted my grading rubrics in those directions. Rather than emphasize the mechanics of writing (more on this below), I allocate most points to categories that fall within what I’ve previously described as “embedding writing within a web of other skills.” That is, I reward students whose writing demonstrates inquiry, comprehension, analysis, argumentation, ideation, empathy, etc.

Then there are the kinds of cases that remind me why we were all so upset when ChatGPT burst on the scene…