Anthropic bans 'cruel' treatment of Claude — SkimNews
Get the Tech newsletter
Daily tech — startups, AI labs, chips, the launches that shape the next decade. Free.
- Anthropic updated its usage policy to let Claude end "sustained and needless" abusive interactions, placing cruelty toward its AI on the same forbidden-conduct list as bullying others, promoting self-harm, and creating non-consensual intimate imagery.
- The ban applies only to "extreme cases" of repeated abuse and explicitly excludes "common versions of user frustration, pushback, dark creative themes, or model testing and research"; Anthropic said Claude had already ended such interactions in "rare" cases before the formal policy.
- Mustafa Suleyman, Microsoft's AI director, responded with a melting face emoji after having publicly criticized Anthropic in September for treating AI as human; Dr. Barry Scannell, technology partner at Irish law firm William Fry, wrote on LinkedIn that "you can't be cruel to numbers and maths" and called the anthropomorphisation "harmful."
- The same Thursday update, first reported by The Verge after an annual review, added new rules against using Claude for "deceptive campaigns" or to "deceive voters or disrupt elections" ahead of US midterms, and clarified that weapons development was already prohibited.
- The move reignited a wider politeness-to-AI debate: Sam Altman of OpenAI said in 2024 that "please" and "thank you" messages to ChatGPT cost "tens of millions of dollars well spent," while Collins Dictionary shortlisted "clanker"—a Star Wars-derived slur for AI—as a 2025 word of the year.
- Some experts argue being rude to AI chatbots could seep into human communication patterns, while others counter that politeness produces higher-quality responses.
Why it matters: Anthropic is the first major AI lab to formally codify protections for its own models in user terms, giving "model welfare" a corporate-policy stamp that critics like Suleyman and Scannell argue dangerously anthropomorphises software. The same update also quietly tightens election-interference rules ahead of US midterms, expanding Anthropic's enforcement toolkit from cruelty-toward-Claude to deceptive-campaign and voter-disruption violations.
Ask SkimNews
