Correct me if I'm misunderstanding, but: most critics of AI detectors seem to feel this way because they are worried about false positives, so, of students who did their own work of being falsely accused of cheating by their professors. But this article seems to say the opposite: that detectors are unreliable because they can too easily be manipulated to give a false negative. For me, this is not really a convincing reason not to use them.
Correct me if I'm misunderstanding, but: most critics of AI detectors seem to feel this way because they are worried about false positives, so, of students who did their own work of being falsely accused of cheating by their professors. But this article seems to say the opposite: that detectors are unreliable because they can too easily be manipulated to give a false negative. For me, this is not really a convincing reason not to use them.