top of page


ChatGPT Detection

Rachel Buttry, Derek Kielty, dora puljic, jack Arbunich, Rajat Majumder


We have built a natural language processing model that can detect whether a corpus of text was created by a human/humans or ChatGPT. We use a dataset constructed from essays written by 6th to 12th graders and ChatGPT essays generated from the same topics. Combining a deep learning model (LSTM) with metrics used in natural language processing analyses, this model can be used as a way to detect if text has been generated from ChatGPT with high accuracy and precision.

Screen Shot 2022-06-03 at 11.31.35 AM.png
github URL
bottom of page