@OP: Proofreading AI

Chatgpt is an AI designed to create text like a human, not to write code. Broadly speaking it's averaging the text it was trained with.

That's like a lazy student randomly mixing and copying the homework of several others in a well sounding form, without understanding what (s)he does.

(Or imagine a politician on steroids speaking a lot of nice sound bits without saying anything.)

In code this might lead to very stupid errors which are hard to spot for humans. Like mixing concepts and variables from different sources.

I hope you don't expect us to proofread the output of your AI. That's your business. Our goal is to teach, not to reward laziness.

Probably one day an AI will evolve to a state where a "nifty" designer can train it to produce stable and maintainable code.

But that's not the same thing, like asking chatgpt to spit out code and asking us to maintain it.