OpenAI shut down the AI text detector due to its low accuracy

OpenAI has shut down a tool that was supposed to distinguish human text from artificial intelligence-generated text due to poor accuracy.

Here's What We Know

The company discontinued the artificial intelligence classifier effective 20 July. The developers admitted that their classifier was never accurate for detecting texts generated by large language models. The company warned of false positives that identified human text as written by AI.

At the same time, OpenAI announced that it is working on new mechanisms that will allow users to identify AI-generated text, audio, images and other content.

There are no details about the new AI content detection tools, nor is there a timeline for when they will be available.

Go Deeper:

• AI detection software discriminates against non-native English speakers - study
• FTC will investigate OpenAI, the creator of the popular chatbot ChatGPT

Source: The Verge