ChatGPT creator rolls out new tool 'Classifier' to trace AI-generated text

OpenAI, the company that created ChatGPT and DALL-E, has now released a free tool that can be used to "distinguish between text written by a human and text written by AIs."

In a press release by OpenAI, the company mentioned that the tool named classifier is "not fully reliable" and "should not be used as a primary decision-making tool." Notably, the latest tool can be used to check whether someone is trying to present the generated text as something that was written by a person, reported The Verge.

What is Classifier? All you need to know about OpenAI's new tool

As per the company, Classifier is relatively simple and easy to use, as the user just needs to have a free OpenAI account to access it. Users are required to just paste the text into a box, click on the button, and it'll show whether it thinks the text is unlikely, unlikely, unclear, possibly, or likely AI-generated.

The company also said in its press release that the tool was trained using "pairs of human-written text and AI-written text on the same topic." OpenAI also issued a warning on the usage of the tool and said the tool isn't always accurate; it can mislabel both AI-generated and human-written text. As per the information shared by the classifier's creator company, the tool only functions after a minimum of 1,000 characters, which is approximately 150 to 250 words, are put into it, and AI-generated text can be edited easily to evade the classifier, the release said.

The release further mentioned that the tool may sometimes "incorrectly but confidently" label human-written text as being from an AI, especially if it’s very different from anything in the training data. It makes it clear that the classifier is still very much a "work in progress."

OpenAI's ChatGPT clears US Medical Licensing Exam

Notably, this isn't the first time that OpenAI is gaining traction for its invention. The company recently announced that its ChatGPT cleared the US Medical Licensing Examination (USMLE). Earlier, the world was surprised when the text-based platform ChatGPT cracked some law and graduate-level exams, but the clearance of the USMLE, which is a Wharton Business School exam for the final test of the MBA programme's operations management course, and four University of Minnesota Law School exams in constitutional law has now caused ripples in the tech world.

Earlier, Chat GPT had helped author a novella within a weekend, which was also sold on Amazon. The clearance of the US Medical Licensing Exam is required for any doctor to obtain a licence to practise medicine in the United States, and the Artificial Intelligence Tool has made it possible by gaining 50 percent scores.

Image: AP