A couple of months ago, I asked a colleague to run an experiment for me. I asked her to give her students assignments and take-home tests with the regular admonitions about academic honesty but no real controls or checks. She then graded them as usual. At the end of the semester, she gave her students an in-class, supervised test on exactly those topics. I am sure you can see what's coming.
With the self-directed and self-supervised tasks, some of her students did as well as expected, but the majority, well over half, showed significant improvement. Some went from Ds to B+, Bs and Cs went to As. My colleague was clearly an exceptional teacher.
But.
In the in-class, supervised test, scores went back to the expected levels with some of the Ds now scoring below a D, the lowest grade available.
After she had the results, she sat down with the class and asked what had happened. She then showed them the scores, both what she predicted and what they produced.
Slowly, most of the students admitted they had used AI for all, or nearly all, of the at-home tasks. Almost all used Chat GPT or Microsoft CoPilot. They did not see this as academic dishonesty or cheating, which is in itself concerning and deserves further discussion.
What stood out was that all felt that grades matter more than learning, or in other words, the only thing that matters is the grade. Grades are not an indication of what they know or can do. Grades have no association with what they know and can do. Therefore, anything and everything to get a grade, no matter how inaccurate and non-representational they may be, including using AI, is justifiable.
So, no AI?
**Please leave your comments and queries below.**
No comments :
Post a Comment