OpenAI has Recently released a rating which aims to determine whether a piece of text has been written by AI platforms such as its own ChatGPT tool.
The company released the tool after various educational institutions and school districts banned ChatGPT because some students were relying on it to write and pass their papers on their own, which Of course there is cheating. Currently, ChatGPT is banned in New York City. Seattle, Los Angeles, and Baltimore public school districts. In some universities France And India Access to the tool has also been restricted. Finally, some states in Australia has blocked students from accessing ChatGPT on school internet networks..
OpenAI describes its text classifier as “a finely tuned GPT model that infers that a piece of text has been generated by AI from a variety of sources, such as ChatGPT.” Despite this claim, however, the company itself admits that the device is unreliable. In its evaluation of English text, the classifier correctly identified only 26% of AI-written text as possibly written by AI, and 9% of human-written text as authored by AI. Considered as What’s more, OpenAI says the classifier may be unreliable on texts that are less than 1,000 characters long and written in languages other than English.
In our own testing, OpenAI’s rating deemed the majority of articles published on Nevin to be “very unlikely” to be generated by AI. However, the tool was indecisive with our recent Nothing Phone (2) coverage, saying it’s “not clear if it’s AI-generated.” When tested with content generated by ChatGPT, the classifier seems a bit dubious, saying the content is “possibly AI-generated”.
This is probably why OpenAI says that the results produced by a classifier should not be the “sole piece of evidence” when determining whether a piece of content was authored by AI. Thankfully, there are other tools you can use. For example, Stanford researchers recently introduced DetectGPT, which helps academics detect AI-generated papers. Also, a computer science student at Princeton has developed a similar tool that can “quickly and efficiently” determine whether an article was created by ChatGPT.
