I've heard some students are concerned that any text submitted to an AI-detector is automatically added to training sets and therefore will eventually will be flagged as AI.