In this repo, you’ll find a dataset of ChatGPT pseudo-news articles generated from the same reader queries responded to in the New York Times’ “Ask the Newsroom” and the Washington Post’s “You Asked”, “Ask Amy” columns (the original journalist-written text is in the dataset too).
The code runs analyses on textual differences between ChatGPT vs journalist-written column text.