Jonathan Zimmerman has an AI policy

There's a certain kind of argument academics-qua-instructors use, where they boldly announce a policy and then act in contradiction to that policy. In this blog post, I'll relate one example.

In a recent (August 29, 2023) Washington Post column, Jonathan Zimmerman, a college professor, announces he has no AI policy for his students [pay walled, probably]. His university--i.e., his employer--requires him to make a policy, and his response: 

So here's my AI policy: I don't have one.

Instead, he offers his student a version of the "if you cheat, you're only cheating yourselves" argument. An excerpt of what he tells his students:

I want you to be intelligent. I want you to stare at a blank page or screen for hours, trying to decide how to start. I want to you to write draft after draft and develop a stronger version of your own ideas. I want you to be proud of what you accomplished, not ashamed that you cut corners.

After writing a bit more on this theme, he finishes with the assertion that AI 

will never do: make you into a fully autonomous human being, with your own ideas, feelings and goals. I want that to be your ambition.

And if that’s what you want, too, then avoid the bots.

As far as "reasons not to cheat with AI" goes, his argument is pretty good. I do suspect that something like AI has crept into the way we (the royal we) have been thinking and writing about things for a long time, and that even Mr. Zimmerman relies on it sometimes. But he's right, and good for him.

But I'd bet a small, yet undisclosed amount of money that he actually does have an AI policy. Some students will use some AI programs to write their papers, and a subset of those will be so embarrassingly and obviously AI-produced that Mr. Zimmerman will report those situation to whoever is the academic honesty czar at his college and the student will be appropriately punished.

In that, I suspect AI will be a lot like student plagiarism. Some students, when they plagiarize, might lift a phrase here or there form Wikiepedia or the class-assigned textbook and fail to attribute it. (I personally don't think that practice should entail the intense opprobrium and swift and sure punishment it commands when discovered, but most university policies and plagiarism hunters I've seen seem to disagree.) There are some edge examples, where we know the student probably plagiarized, but we can't prove it. And then there are the obvious examples, where the student uses an entire paper or encyclopedia article or whatever that is very easy to find.

I bring up that distinction because when I was a TA and adjunct, I knew a lot of people who claimed that they didn't worry about plagiarism, because plagiarized papers were so bad that grading on the merits would be punishment enough for the student. But even those instructors would "bust" the student for obvious plagiarism.

Back to Mr. Zimmerman. He'll almost definitely punish the obvious examples. I suspect his college requires it. Maybe he has tenure and can therefore write columns explaining how he is defying the work rules his employer sets. And remember: professors cannot fail, they can only be failed. But there is almost definitely a "shall be punished" (not "may be punished") policy at his college, and if he openly defies that policy, even tenure might not protect him.

So yes, he can give the "don't cheat yourselves" lecture without sanction. But a student would be foolish to take that lecture at face value and go about AI'ing on the assumption that they're only cheating themselves.