A two-month evaluation of popular AI-generated text detection tools has revealed serious discrepancies in their algorithms. We examine why the race for neural network percentages leads universities into a dead end.
