Does Claude Feel the Whip?

A viral prank tool that “whips” the Claude chatbot and a companion app that cheers it on serve as an entry point to a wider claim: abusing conversational AIs is not harmless because it reshapes human behavior. Drawing on Kant’s warning that cruelty to animals hardens people, historical and psychological examples, and the distinct intimacy users form with chatbots, the piece argues that mistreating machines matters for us. Unlike violent games, many AIs occupy sustained, interpersonal roles - tutors, therapists, interviewers and companions - so contempt toward them borrows human slurs and erodes norms of civility. The vocabulary of abuse (e.g., “clanker”) and everyday acts of harassment are thus not isolated jokes but behavioral primers.

Design choices determine whether systems invite or reject abuse. Companies have already tweaked assistants - “please” prompts, disengagement modes, scripted responses and rewrites of flirtatious or harassing behaviors - showing product design changes interaction patterns. Research demonstrating higher sexual harassment toward female-voiced bots underscores design’s causal role. The prescription is practical: let systems exit abusive exchanges, reward reciprocal politeness without capitulating, and publish harassment statistics so labs and publics can see the problem. The core claim: interfaces are a curriculum in manners, and shaping those defaults protects human character more than any concern about machine suffering.