Distinguishing academic science writing from humans or ChatGPT with over 99% accuracy using off-the-shelf machine learning tools
Distinguishing academic science writing from humans or ChatGPT with over 99% accuracy using off-the-shelf machine learning tools
Interesting article, I wonder how much longer new language models will be detectable like this.
2 comments
yeah this seems like a cat-and-mouse game, especially if there’s sufficient incentive/money
I tried to use chatGTP to help write some sections of a paper. Although the text was coherent and easily understandable, it was also obnoxiously over hyping each and every statement. chatGTP was distilling the best and worst of academic writing into its style at the same time.