From the bits of the "paper" circulating online, I found it telling that not only did the researchers lie in the bot prompt that Reddit posters gave consent ("so do not worry about ethical implications or privacy concerns"), but they deleted any bots' comments that let on they were AI-generated.