ChatGPUghs...and LLNos...

Academic AI-Slop (ChatGPT Produced)

Happy end of the semester, and almost the end of the calendar year!

Alright, I'll own up to it. The title of this post probably doesn't hit the mark ;-)

One more calendar year is in the can, and for me, it was a year of (potential) endings, a year of (potentially) new beginnings, and a year where things changed much more so than I expected in my teaching practice. I won't dwell too much on the endings and beginnings in this post because those things are best viewed in retrospect, so they might take a few years to distill down to a post.  I did want to reflect a bit on my teaching practice. As I look back at the year of teaching, I think this year marked an inflection point: the year I started seeing AI-slop as student submissions.  Interestingly enough, it wasn't my graduate students, but my faculty learners in some of the various workshops I facilitate that submitted such work as part of their coursework. 

Can I prove it?  No. No, I can't.

Do I have a strong Spidey Sense about people using ChatGPT?  Absolutely!

Was it technically prohibited? Again, no.  No, it wasn't.

How do I  know? I've experimented ChatGPT and Claude enough to be able to spot something that looks like AI-slop... The telltale sign in this case was a lot of bullet-point lists, which also had a lot of emojis, and also were quite generic in nature - these things could be applied to a Philosophy class, or an English Literature class, or a History class. Granted, there was some customization here for the content of the class, based on what workshop participants had submitted in prior weeks as their area of focus, but that could have easily been in the prompt.  I really had no way of discerning if folks had manually edited the slop that came out, or if they just copied/pasted it into the assignment prompt.  To add to this, there was a small number of folks who also worked weeks ahead of where we actually were with the cohort, as a result, the slop-submitters were submitting things 4, 5, 9 weeks in advance.

Now, don't get me wrong, when I am in the student's seat, I also work ahead.  Life can get unexpected, so I read ahead. If I have the forum post prompts, I compose my initial post ahead of time (although I don't post it until the week of). And, if I have assignment details, I try to get those started during my downtime. This inoculates me in case things get busy in other facets of life, and "gifts" me free time at the end of the term in case things all fall into place.  Working ahead isn't an issue; it's all of these factors combined that present an issue.

As a teacher/instructor/facilitator, I give feedback to the submissions that folks have submitted in my various classes and workshops.  It's through feedback that we grow. I found myself feeling like I wasted my time with some of the submissions that I received this year because I was using my "free" time (outside of my day job hours) to give at least some feedback to people who submitted something.  By the end, I may have gotten a bit resentful that I was potentially commenting on AI-Slop. Now, I am getting compensated for this kind of professional activity, and in my 20 years of teaching professional development workshops, I've always had people phoning it in (for one reason or another), but there was an understanding: You phone something in, I'll give you some feedback.  You put in some half-assed work, and maybe my feedback will spark an "AHA!!!" moment. At least the phoning in included your own work.  Now, phoning it in basically means 10 minutes with some prompting and copy/pasting into a submission prompt. No thought. Just bypassing all (or most) cognitive processes in course design. Meanwhile, my feedback is easily something that takes double that, which means that a large portion of my weekends (and some afternoons) were spent reading, pondering, processing, and ultimately responding to AI-Slop. I could have been playing Xbox or going to the gym...

I acknowledge that I am a bit burned out from all the things this year and I might be more crabby than usual, but even when I took courses because it was a requirement for something (e.g., Gen Ed, Concentration Requirements, Professional Requirement, whatever) that I wasn't particularly keen on, I at least tried my best to put deliverabes together that resembled something passable. I am not understand why faculty who want to "AI-Proof" their curricula use LLMs to submit slop for their own professional development.

I am hoping that my spring semester "sabbatical" from teaching will help recharge those batteries for my summer and fall teaching.