College and high school students have turned to AI software such as ChatGPT to automate and expedite the often tedious process of writing essays. This naturally poses a major problem for teachers, and has led to a cat-and-mouse game between increasingly advanced content-generation software and efforts to detect automated content.
Here are some solutions I devised that does not involve software:
1. Rely more on class participation, quizzes, and in-class assignments and less on essays. This also helps to assess baseline competence for each student, so cheating becomes more obvious. A student who turns in an “A-grade” essay but does not contribute, demonstrates poor understanding in class, or performs poorly on quizzes would be a red flag.
2. At the first day of class, everyone must write a short in-class essay, without the help of smartphones or computers, to assess baseline competency. Assigned essays are then compared to this assessment. Anyone whose writing ability is too advanced relative to his or her assessment would be flagged for additional review.
3. Breaking long-term assignments into stages, with reviews at each stage. So instead of a term paper being due at the end of the semester without any prior review, each student must submit periodic reviews at impromptu notice to track progress along the way. Failure to produce an essay or if the final essay is drastically different, would also be a red flag. This is similar to surprise drug testing in sports.
The major problem with these solutions is they entail a greater workload for the teacher. At the college level it’s assumed that teachers are not supposed to micromanage students.
Students can also get around this by modifying the settings to write intentionally dumbed-down essays, with deliberate grammatical mistakes and common misspellings. If the objective is to only pass the class, then writing C-grade essays is good enough and lessens the risk of detection. I can see a market forming in which students pay freelancers to modify AI-generated essays to appear more human-like or closer to baseline ability, with certain idiosyncrasies that are unique to the student.
But also, teachers have some responsibility to make the coursework engaging enough that students do not feel as compelled to cheat, but credentialism, again, is a recurring problem. Students cheat because they perceive, correctly, that the coursework is a waste of time and inapplicable to their desired profession or major. Others cheat because they assume, also correctly, that in a work environment the cost of incompetence is distributed among the entire firm and that there are ways to hide or minimize this (fake it until you make it), as opposed to the classroom setting in which incompetence is more apparent.