Meeting KI-Postdocs: 'Artificial intelligence outperforms human'

Full Title: "When artificial intelligence outperforms human social cognition"
(Dr. Uwe Peters)

Suppose an artificial intelligence (AI) excels us in social cognition in that it ascribes
psychological characteristics to people more accurately than us. It seems epistemically rational to
then set aside our own judgments on those characteristics and defer to the AI. After all, being more
accurate about people is better than being less accurate. I argue that this intuitive view is
epistemically and ethically problematic.