Schools need clear rules on AI detection tools

My daughter's university just ran a batch of essays through an AI detector and flagged three students, one of whom is now facing a misconduct hearing. She wrote her essay in Google Docs with version history intact, so she can at least prove her process, but plenty of students can't. The detector vendors themselves admit false positive rates that would be unacceptable in any other assessment tool. If institutions are going to use these systems to accuse people, they need published accuracy figures, a proper appeal route, and a presumption that the tool is wrong until proven otherwise. Otherwise we're teaching students to fear writing.