Generative A.I. is Ruining Online Classes
I love learning, and virtual courses increase accessibility, but A.I. is enshittifying them
I love school
If you’ve been reading my work over the past two years, you’ll know I returned to University to upgrade my degree(s). I live rurally and also work from home, so online classes are a fantastic opportunity for me to learn while continuing my advocacy work. Unfortunately I’ve noticed a very frustrating trend which is degrading the quality of instruction, learning, and the overall virtual classroom experience.
Let me be clear: I am not blaming the professors or my fellow students whatsoever. This is a systemic issue that is bleeding into just about every facet of our lives, and it’s causing specific problems for Universities which I’ve seen worsening since I resumed my studies in 2024.
I’ll start with the obvious, but very briefly, since we all know about students submitting papers written by ChatGPT and other generative artificial “intelligence” (A.I.) programs. This on its own has led to professors assigning fewer papers, using more exams and quizzes for assessment instead, and when they’re done online this generally means multiple choice tests.
I. Hate. Multiple. Choice. Tests.
Yes, I’m Middle-Aged
It’s been a while since I took an educational psychology course (and by a while, I mean nearly 20 years), but I most certainly remember learning in various classes that multiple choice (M.C.) questions are only good at measuring recall and do not assess for deep conceptual comprehension. Want to know what happened when I went searching for updated literature on the subject? I got more than two pages of results on using language learning models (LLMs), ChatGPT, and other A.I. programs for developing M.C. questions. Ffs.
Like most AuDHDers, I have poor working memory. That means I struggle with short-term recall and I really suck at rote memorization. In order for me to actually learn and remember information, I need to understand it, and be able to apply that knowledge to real life. That’s why long-form assessments like essays and papers are ideal for me. I can demonstrate my understanding of the subject matter, and explore it in a complex way that is interesting and requires deeper thought than merely identifying the correct answer out of a line-up.
Additionally, I often find M.C. questions confusing. This may be my Autistic literal thinking and/or pragmatic language use, but I could usually make a strong argument for at least two of the options in nearly every M.C. question I’ve come across. I understand that’s kind of the point, that usually two answers look like they could be correct, and purposely requiring a certain level of knowledge to discern between them and pick out the right one. I’m not talking about that, I’m referring to questions which assume an introductory level of knowledge (as most undergraduate courses do), and oversimplify the concept for the purposes of testing. Also some questions are just poorly written, confusing, and vague.
Lastly, M.C. questions leave no room for demonstrating knowledge despite not getting full marks. With long-form questions, a student can offer explanations and arguments based on the coursework to support their answer. Even if they may have gotten the answer wrong on a M.C. test, they can still show their comprehension and application of the concepts learned in the class, something which is impossible to do with M.C.
That’s not even the worst of it
That’s not all, oh no, that is not all.
