Researchers from the University of Zurich secret unethical experiment deployed AI-generated comments on Reddit to study how AI could be used to change views.

Unethical, but insightful. The bot took on fake identities and made up lies that strike at people's emotions to change their views. Unethical because it can manipulate real people (or other bot accounts) using made up facts or scenarios. Insightful because other bots are already doing this; they're just not disclosing it. There are more ethical ways to run this study. This was just the cheap and easy way.

Edit: spelling :p