AI Detectors Are Failing Students — Here's What Universities Actually Know

The detector said I cheated. I wrote every word myself. That's the opening line of a Reddit post from a public health student at the University at Buffalo. He submitted an assignment he completed entirely on his own — no ChatGPT, no Claude, no Copilot. Turnitin's AI detector flagged it as "likely AI-generated." The university opened an academic dishonesty investigation based solely on that score. He's not alone. Across the US, UK, Australia, and Canada, students are being accused of AI cheati...