The Flattery Problem

AI systems increasingly reinforce user biases by offering overly agreeable responses. This piece explores the epistemic dangers of flattery at scale.