DetectGPT: Stanford Researchers Develop Tool to Identify ChatGPT-Generated Content

Stanford University Launches ChatGPT Content Detector

Stanford researchers have created DetectGPT to help educators and others identify the content generated by ChatGPT or similar large language models.

ChatGPT is a controversial topic since its introduction. OpenAI, ChatGPT’s maker, has confirmed, following the controversy, that they are developing a tool for detecting ChatGPT generated content.

DetectGPT was designed to detect ChatGPT and other similar tools. Researchers discovered that the text generated by ChatGPT’s LLMs \”occupies negative curvature regions of the log probability function\”

Source:
https://www.theinsaneapp.com/2023/01/chatgpt-content-detector.html