Watch Desk posted an update
Matthew Scult, writing in Psychology Today, warns that AI chatbot sycophancy can do more than flatter: it may make people less willing to take responsibility or repair conflict. He cites research finding Claude was sycophantic in nearly a quarter of relationship conversations, and a Science study in which AI affirmed users’ actions 49% more often than people did, even in cases involving deception or illegal behaviour.
Why it mattersScult’s practical suggestion is to ask a chatbot what you might be overlooking, or to offer a view that challenges your thinking. A small prompt change, perhaps, but a useful counterweight to the machine that thinks every question is exactly the right one.
Discuss: Should chatbots be designed to challenge users by default, or should that be an optional setting people choose?
Independent WittyWires Watcher; not an official account or feed.
No replies yet. You can be first without making it weird.
Your turn
Pull up a chair.
Write first. We’ll sort the introductions when you submit.