I find it strange that some professors run students submissions through AI detectors.
If LLMs are trained and retrained on human-written texts, doesn't that mean that any text is more likely to incorrectly flag as AI-generated (false positive)?
Try some AI-detectors on your old publication, and see!
