Stanford, 1 February 2023 (GNP): People use ChatGPT, an artificial intelligence (AI) chatbot created by OpenAI, for a variety of purposes, from answering chance inquiries to having the chatbot complete tasks for them.
A programme called DetectGPT, created by Stanford researchers will assist teachers in spotting content produced by ChatGPT or other comparable large language models (LLMs).
Concerns regarding students using ChatGPT to complete their papers have existed since the chatbot’s launch. The developer of ChatGPT, OpenAI, acknowledged that it is developing a method to identify information produced by ChatGPT after New York City public schools forbade pupils from using it.
Also Read: PTCL interested in buying Telenor INC
A tool called DetectGPT has been created to identify content produced by ChatGPT and other tools of a similar nature. The text produced by LLMs like ChatGPT “occupy negative curvature regions of the model’s log probability function,” researchers claim. The researchers have developed a new standard for determining if a piece was produced using an LLM utilizing.
This enables DetectGPT to determine whether the given passage was produced using a certain LLM using log probability. The accuracy of this app, according to the researchers, is 95%.
