Anthropic Bans Cruelty to Claude, Still Won't Say What It Protects
ojosilva · hn · 2026-10-11
Anthropic's updated usage policy now includes a ban on "cruelty to Claude," marking one of the first times a major AI lab has written model-welfare-adjacent language into its terms.
- The company hasn't explained what the clause actually protects — whether it reflects an interest of the model itself or simply shapes user behavior norms.
- Critics argue Anthropic is nudging model welfare into official policy while refusing to disclose its reasoning, calling for more transparency.
More from Companies & People
- Anthropic reportedly asked AWS to urgently pull Sonnet 3.5, 3.6 and 3.7 from Bedrock — repligate · 2026-10-11
- Researchers blast multi-site AI conferences: splitting venues kills serendipity — iScienceLuvr · 2026-10-11
- Toby Walsh's The Shortest History of AI earns Kirkus praise ahead of Nov 4 release — TobyWalsh · 2026-10-11
- Valar Atomics CEO on building a culture of transparency even when it hurts — ctjlewis · 2026-10-11
- Palmer Luckey slams ZeniMax lawyers for deposing his 84-year-old neighbor in Oculus lawsuit — ctjlewis · 2026-10-11
- Delip Rao mocks big tech CEOs' sudden rush to signal they are 'team superintelligence' — deliprao · 2026-10-11