How To Detect If Chat GPT Wrote An Essay (Plagiarism Checker) - GPT Zero

In this video, the speaker is introducing GP2 zero, a new tool designed to detect and protect against plagiarism and academic dishonesty, specifically when it comes to using Chat GPT. The speaker demonstrates how to use the tool by having Chat GPT (a language model) write a 3000 word essay on the topic of whether or not Hamlet loves Ophelia.

The speaker then copies and pastes the text into GP2 zero, which then checks for perplexity and bursting. Perplexity is a measure of how well a language model is able to predict the next word in a sequence and the lower the perplexity, the more likely the text was generated by a language model. Bursting checks for repeated phrases and the higher the bursting, the more likely the text is generated by a language model.

In this case, the tool indicates that the text has a low perplexity and a high burst, indicating that it is likely to have been generated by a language model. The speaker advises that while GP2 zero can be useful as a tool to detect plagiarism and academic dishonesty, it is not a substitute for independent thought and writing.