OpenAI Abandons AI-Crafted Text Detector Due to Pitiful Accuracy Rate

OpenAI has discontinued its AI classifier, a tool designed to detect if a text passage was written by another AI, due to its low accuracy rate. There was a widespread notion that AI could reliably detect AI-written text, however, the rapid developments in language models made it challenging to detect potential AI “tells.” A test conducted by TechCrunch revealed the unreliability of such tools, with OpenAI’s classifier correctly identifying only one out of seven AI-generated text snippets. Despite its limitations, some users trusted the classifier tool to identify AI-generated content. OpenAI joins other firms in a White House-led commitment to ethical AI development, including robust watermarking and detection methods, although no company has yet demonstrated a foolproof method.