It’s not. I work in academia. Dumb, illiterate students think they’re sneaky and they’re not. The only way to get in trouble for using AI is to use AI.

Identifying LLM writing is extremely trivial, independent of the fact that detectors have low false positive rates.