I get some of the outrage about this happening without users knowledge.
but humans still oversaw every post. and I can understand why researchers didn’t want to announce this study in advance.
the results suggest a present day where immoral, powerful people can persuade others of their views. I don’t think influence operations will be asking for informed consent either!
I guess AI-enabled persuasion isn't exactly new. but the effectiveness of these personalized AI bots relative to human baseline is making me internalize what this means for our near future.
we need to develop a stronger societal immune response to persuasion bots
• • •
Missing some Tweet in this thread? You can try to
force a refresh