I’ve seen a lot of these articles now and I’m increasingly convinced the answer is just that teachers and profs are going to have to go back to the old methods of teaching, even tho they suck. Too many profs assign easy homework and exam style evaluations because students like them and they’re easy to grade. But that stuff has become very easy to crack, even before AIs. And the solutions are draconian, if you do an online exam some profs will track your eye movements (!!!) to see if you’re flipping tabs in a browser. What a joke.

The solution then is to go back to the traditional essay, and then ask a question that only an engaged student can answer. An AI might be able to scrape google and figure out who John Kennedy is, but if I ask a student to “Identify the long 1963 thesis and analyze its impact on the American decision to go to war in Vietnam” I don’t know that AI, at least for now, is going to be able to answer that. Moreover, the real secret is that I control all the information presented. Student essays that start talking about subjects outside the question, indeed outside the material presented in the course, begin to look suspicious.

Students don’t like essays because they’re hard, and profs don’t like grading them because it takes forever. But it’s the way to go. And IMHO long form assessments more closely resemble the kind of work most people are going to do, esp for humanities skills.