A program created by OpenAI to differentiate between human and artificial writing has been withdrawn due to poor accuracy. In a (updated) blog post, OpenAI stated their intention to stop using its AI classifier as of July 20. As the company put it, "We are working to incorporate feedback and are currently researching more useful provenance techniques for text."
OpenAI announced that it plans to "develop and deploy mechanisms that enable users to understand if audio or visual content is AI-generated" while it discontinues the tool to identify text generated by AI. However, it is currently uncertain what these procedures would entail.
OpenAI warned that the classifier could create false positives or human-written words and publicly admitted that the classifier had never been very good at detecting AI-generated content.