The makers of ChatGPT are releasing an AI detector for schools to prevent cheating

For months, professors, teachers and school administrators have fretted over the prospect of students handing in work produced by AI chatbots, calling it an existential threat to traditional education.<br><br>OpenAI is now launching a tool to solve this very problem. The company is releasing a new system designed to detect whether text was written by a human or a machine. The new service is called AI Text Classifier.<br><br>“It’s designed primarily for teachers and educators to get a sense of whether the text looks like it was written by AI or not,” said Zachel Ilk, a spokesperson for OpenAI. “This classifier gives an extremely high probability result that you can trust. If you get a high probability result, you can trust that it’s written by AI. If you get a low probability result, you can’t trust it. There’s a third result that’s ‘uncertain’ and that’s for example when ChatGPT was only used for a part of the text.”