General Discussion

As a teacher, you can pretty much use whatever tools you want to grade students' work - and I know some who are using AI as a way of doing a first cut. Their contention is that it frees them to focus on more important aspects of the work that AI is not yet good at (picking out made-up crap, appreciating the flow of an argument among essays which each received the same number of points for simply mentioning required elements, etc.)

I, personally, don't think it is sophisticated enough to be a useful tool yet, and I'm not aware of any LLM AI which is exclusively trained on licensed content so there are ethical issues. So I wouldn't use it.

But some instructors are already using it for grading. And - given the increasing work load (increases in class size/decreases in number of faculty) - anything which allows instructors to focus their time on things which AI can't do, or can't do well, isn't inherently bad. I have always spent 80-100 hours a week for the 20 years I taught (11 in high school, and 9+ in law school). As a salaried employee I was never paid for the 40-60 hours I was donating. I automated as much of my gracing work as I could so that I could spend my time on things that were formative, rather than evaluative. Most teachers I know resorted to multiple choice tests (which is the fastest - but least accurate way of evaluating knowledge, with no formative feedback), fewer evaluations, etc. simply to get by without putting in the time required to complete the actual workload.

As a student (which I am currently), I would rather have a teacher who figured out how to spend more of their time creating meaningful lessons and being available to students than grading. And as class sizes increase, the ratio of grading to teaching also increases. Something has to give.