A university provost's own op-eds got flagged as AI-written by a detector marketed as near-zero error, adding to a run of false positives serious enough that Curtin University switched its AI-writing detector off entirely rather than keep penalizing students for it.