OpenAI creates a new classifier to detect AI written text

OpenAI has created a new classifier to detect text written with Artificial Intelligence (AI), the company said on its website Tuesday, where the service can also be accessed free of charge.

While it is impossible to absolutely detect all AI-written text, OpenAI believes that the development of classifiers can serve to mitigate “false claims that AI-generated text was written by a human.” The fact would not be of minor importance, considering its potential use in dealing with automated disinformation campaigns, as well as academic frauds, to give just two examples.

“We’re making this classifier publicly available to get feedback on whether imperfect tools like this one are useful. Our work on the detection of AI-generated text will continue, and we hope to share improved methods in the future,” OpenAI stated in the release.

In tests on a “challenge set” of English texts, the tool correctly identified 26% of AI-written texts (true positives) as “probably AI-written,” while, on the other side, it incorrectly flagged 9% of human-written texts as AI-written (false positives).

“Our classifier’s reliability typically improves as the length of the input text increases,” the company added. However, compared to the “previously released classifier, this new classifier is significantly more reliable.”