Researchers at the University of Zurich wanted to see how persuasive AI could be — but did not get permission to do research on the actual humans on Reddit that the AI was interacting with. There were ...

A group of researchers covertly ran a months-long "unauthorized" experiment in one of Reddit’s most popular communities using AI-generated comments to test the persuasiveness of large language models.