Yes, agreed. Actually the very existence of the problem described in the original post is an indication that the work being assigned is ultimately useless. If t

Yes, agreed. Actually the very existence of the problem described in the original post is an indication that the work being assigned is ultimately useless. If the teacher can't reliably distinguish between a student's answer and an AI answer then what's the point of giving that assignment? Kids with street smarts or a hacker mentality are always going to figure out the easiest path from A to B, and it's arguable if we should even be discouraging that—it's quite an important skill in real life. We're at a major transition point and I don't envy teachers who need to figure out ways to adjust to it, but giving the same exact assignments clearly isn't going to cut it. They should instead try to come up with assignments where an AI may be a helpful assistant but isn't able to complete 100% of the task to a high standard. Then evaluate the result regardless of how the student got there. Presentations, discussion-based evaluation, and projects with a physical component are all examples. I remember when I was a kid, some teachers would assign research papers and then tell us we weren't allowed to use the internet; we needed to go to the school library and use physical books instead. This holy war of teachers vs. chatgpt is likely going to seem similar in hindsight.