Professors struggle to detect AI-written student essays. Early AI detectors were inaccurate. Turnitin now flags AI in nearly half of US student submissions. Turnitin claims a 1% false positive rate and 15% false negative rate. Vanderbilt disabled Turnitin’s detector over false positive concerns. Pangram claims a false positive rate of one in 25,000. Some universities discourage or ban AI detectors. Students use humanizer tools to evade detection. AI detectors can mistakenly flag non-native English speakers. Some educators argue students should use AI responsibly. MIT recommends rethinking education to reduce reliance on detectors. Indiana University bars professors from using AI detectors. Harvard advocates for acceptance of AI in writing classes.

• 750 false positives per year at Vanderbilt
• 1% false positive rate for Turnitin
• 15% false negative rate for Turnitin
• 1 in 25,000 false positive rate claimed by Pangram