Last year I had several blatant examples of students using AI to generate their assignments and I spent a few months figuring out how to create new assignments that were less easily plugged into ChatGPT and would force the students to read and digest the course materials and generate responses that would demonstrate that they put their own thoughts into their assignments.
I am not opposed to the use of AI to help students with things like grammar and syntax. Although I admit that AI has lessened some of my amusement. I no longer get responses like “smartification” instead of “smolitification”. I don’t see sentences like this “Downstream migration with wind speed in the direction of the migration path results in increased migration rates later in the season.” anymore, because AI fixes them. I mean…that’s good, but it’s way less entertaining.
No, I had to move away from essay questions because AI excels at essays.
What to do.
After chatting with a few colleagues in teaching and hearing about how they are combatting the use of AI but embracing it in different ways, I decided to ask AI how to work around AI to ensure that students were having to review and digest the material instead of just outsourcing their papers to a robot. And AI proved quite helpful. It made some suggestions…some I rejected, some I worked with, and I landed on a series of tasks/questions related to the specifics of the material that provided room for individual thinking.
So my assignments this year are very different, asking for specifics that force the students to turn directly to the materials on the pages of the online materials, and then permits them to use external resources to bolster their understanding. They were required to sign off on a use of AI attestation, and attest to what AI resources they used, and how, for each assignment.
It’s not perfect, but after grading for a couple of days, and getting the new rubric sorted out in my own head, it went relatively well and the responses were very much more ….’human’. There are some changes needed, some things that more students misinterpreted than should have been the case, so a few word changes necessary to better clarify expectations and reduce misdirection on my part. There is always an assumption in writing that everyone else will understand your thinking, but that’s rarely the case because the author of anything has thoughts and ideas that never make it to the page. I tell the students that I can only grade what they write, not what they meant to write, and I have to apply the same to myself when marking responses to something that perhaps I didn’t make as clear as I thought I did with respect to what I was looking for.
I did make one grading error on one paper…and towards the last paper I had a niggling feeling I’d marked someone’s table of comparisons of two pathogens out of 20 instead of out of 25, but my tired brain didn’t find it on scanning back a few so I left it and uploaded the grades.
If one thing can be counted on, if an error in grading is made and it is in the wrong direction, the student will let you know ASAP.
The email was in my inbox this morning. I knew I’d done it somewhere, he found it immediately. Problem remediated.
