In an essay for Noema Magazine, Yaniv Regev argues that abusing AI systems like Claude is worth scrutinizing not because it harms the machine, but because of what it does to the person doing it. Drawing on Kant’s argument that cruelty to animals corrodes human character, Regev contends that AI labs are effectively shaping users’ manners through interface design choices. He proposes that systems be allowed to end abusive conversations, that labs reward courtesy without encouraging sycophancy, and that companies publish data on hostility directed at their systems.
