SignalTech/AI
Anthropic Bans Cruelty to Claude, Still Won't Say What It Protects
Anthropic bans cruelty to Claude, highlighting AI safety policy without specifying protected entities.
Context
What happened: Anthropic's 2026 Usage Policy update, announced October 8, bars "sustained and needless abusive or cruel behavior" toward Claude models starting November 12, 2026, exempting frustration, dark creative themes, and model testing.
Key numbers: Effective November 12, 2026.
Why it matters: Enforcement relies on Claude's existing ability to end conversations, a tool Anthropic says most users will never trigger, but no firing-rate figures have been published. Microsoft's Mustafa Suleyman disputes the premise, calling AI "sequence completion engines, internally hollow.
What changed
Anthropic's 2026 Usage Policy update, announced October 8, bars "sustained and needless abusive or cruel behavior" toward Claude models starting November 12, 2026, exempting frustration, dark creative themes, and model testing.
Why it matters
Enforcement relies on Claude's existing ability to end conversations, a tool Anthropic says most users will never trigger, but no firing-rate figures have been published. Microsoft's Mustafa Suleyman disputes the premise, calling AI "sequence completion engines, internally hollow.