OpenAI introduced a new tool, and it’s not ChatGPT

Whereas folks had been busy judging ChatGPT for the nice or the fallacious values it could add to our lives, OpenAI got here with a brand new and smashing device, the AI Classifier. ChatGPT, aside from being within the talks for the incredible issues it will probably do, was additionally getting a infamous tag for the extent of misuse it may encourage.

The AI Classifier is skilled to distinguish between AI-written and human-written textual content. It’s a language mannequin skilled to accurately label AI-written as “possible AI-written” 26% of the time. Whereas it additionally predicts false positives 9% of the time, detecting human-written textual content as AI-written.

OpenAI claims that this classifier will “inform mitigations for false claims {that a} human wrote AI-generated text: for instance, working automated misinformation campaigns, utilizing AI instruments for educational dishonesty, and positioning an AI chatbot as a human.”

The AI device remains to be within the making, and the house owners have been trustworthy about revealing its limitations to the general public. They declare that the applying shouldn’t be the first decision-making device because it has some drawbacks, resembling being unreliable with brief texts. It may carry out higher with languages apart from English and simply predictable texts.

Whereas AI Classifier remains to be a piece in progress, the makers have launched it for public use to examine whether or not they want the device and see how effectively the viewers accepts it.