OpenAI has released a software tool that can identify if text has been generated by artificial intelligence
The company behind the popular chatbot ChatGPT says it has trained a language model to distinguish between something written by a human, and A.I.
‘AI Classifier’ uses a variety of providers to address issues such as automated misinformation campaigns and academic dishonesty.
The detection tool is very unreliable on texts under one thousand characters, and AI-written text can be edited to trick the classifier.
Many schools across the world have already banned ChatGPT from being used for projects.