Honestly, she’s in a better position to “do” something about this than many of the rest of us are… she can certainly put pressure on department heads that report to her, as she’s the dean of the school…
But yeah, I think many of us agree that this is a problem - but so is having a course load so heavy that one can’t really have much time to create interesting course work that helps people think critically about the material. Most people doing survey classes tend towards multple choice quizzes or tests, because that’s what people have the time to create and grade… and is it little wonder that more are going for ready-made packages from publishers, either? It’s just so much easier than re-tooling each semester when you think you can do something better…
The students often see the material as coming between them and the grade, where the teacher sees the grade as coming between the student and the material. So this method of getting the grade without needing to master the material is going to be a part of education, like drugs are a part of athletics. Making the penalties for getting caught stiff enough may deter most students, but not all. People who feel that cheating is ok if it gets them the degree/diploma will cheat. That this form of cheating is harder to detect is a problem. Maybe colleges will have to come up with some more rigorous, in person, no assistance allowed exams to weed out those who relied on software to do the work.
This has me thinking about how we weren’t allowed to use calculators in math classes in high school until we got to a high enough level that we’d shown we understood how to do the work, and at that point freeing up time and mental energy by letting a machine do the basic grunt work was considered the better option moving forward.
And likewise with graphing calculators once we’d shown we could make competent graphs on graph paper.
ChatGPT can write a credible first draft. Honestly, the real work is taking that draft and polishing it to the point that it’s at a master level. In order to do that, you would have to put in the work to really understand it and for some people, I could really see the value in that. For some people, getting to that first draft (or even starting one!) is the hard part. Those same people often really kick into gear once they have a draft to polish.
I dont know that that’s always been the real work-- you have to have something better than a turd to polish.
But yes, if AI can write a workable first draft, then useful analogies seem makeable to other processes that used to require a lot of first steps that we no longer have to make.
So speaking of calculators, I wonder what older math instructors are remembering and thinking about now in terms of parallels with the changes they had to make in response to the ubiquity of a task-doing machine. Forbidding calculators until students learn the basics seems somewhat workable, but the existence and affordability of calculators surely has come to mean working them into one’s teaching in some ways. Students are going to use them no matter what, and students are going to use AI for writing, no matter what. Teachers of courses that require writing may well need to make a similar shift.
Yeah, but a brain is trainable for a few digits of basic arithmetic and trig factoids fairly easily, and wow does that speed things up when you’re pushing equations around on a page, or looking for obvious mistakes. I had a colleague who was pretty good at square roots to 6 digits, but that was a bit of a party trick. If you’re cranking out serious calculations then it’s computer time anyway.
At that point, you need to have done a good job of the symbolic analysis to prove that what you’re computing is sensible and even can be done. That’s the “real” math. If you get that right then, instead of a large, finite number of calculations, you get a reduction or simplification equivalent to doing an infinite number of calculations, which is where pushing the equations around gets really valuable. (I don’t care how many monkeys you have in your GPGPU, it ain’t ∞.)
Side bar: I tell my kids that, as a mathematician, I rarely deal with numbers other than e, i, 2 and very occasionally 3 (1 is implied, 4 is just 2 · 2, 5 only shows up as 0.5 = ½, sloppy… if I see a 7… I’ve done something really wrong…)
I “laughed” because as someone who’s always sucked at math, your comment is nearly all Greek to me.
Are you saying there really is no useful analogy for writing instructors to make between the advent of calculators and that of writing AIs, in terms of changes they should make (or maybe will even have to make) in their teaching?
Fortunately, it takes a lot of practice to get as fast writing equations by computer as by hand, I never saw anyone using equation typesetting in undergrad. At worst, pencil and paper means they have to copy, and that’s usually obvious.
An analogy arises from checking written computer programs. I’d warn the students at the start of term of a couple of algorithmic mistakes that would get them instant 0’s on assignments, the sort of thing you get if the code had been copied from somewhere. Making them write in an uncommon computing language helps too, harder to find sources to crib. Also, coding style is probably more distinct than writing style. After you work with someone for a bit, you can spot their programming style from across the room just by how it sits on the screen or page.
The only thing that comes to mind for me, is that to get certain tasks done, you have to have certain capabilities in your head that you can only push into our mushy, wet excuse for a brain through practice. Enforcing that practice means personal relationships and work on the part of the teacher, which is “expensive” (in quotes because quality education is key to quality work, which is an investment that pays off in the end).
But there are other factors at play, AP notes: University budgets are getting tighter, and institutions are increasingly relying on part-time adjunct faculty — who don’t have the possibility of tenure — to teach classes. And that’s not just a red state issue. “Tenured faculty numbers have been declining even in more liberal states.”
One clear trend in the last 20 years – more women and non-cis-het White men getting tenure. Not as fast or as many as would be good, but enough to scare the White establishment.
