Anthropic's 2026 Usage Policy update: deceptive campaigns, autonomous actions, abuse rules
kimmonismus · x · 2026-10-09
Anthropic published its 2026 Usage Policy update, effective November 12. Key changes: a new consolidated section banning deceptive campaigns and artificial activity, informed by observed misuse of Claude for influence operations (fake account networks and fabricated news sites by state media, propaganda offices, and commercial firms); new examples reflecting emerging misuse in influence operations, weapons development, and surveillance per its latest threat intelligence report; clarified requirements for high-risk health and finance use cases; new controls for autonomous physical actions; and a new provision addressing abusive behavior toward models. Most updates clarify existing rules for Claude's longer, more autonomous work.
Related event: Anthropic bans "abusing" Claude in new usage policy, sparking backlash(21 posts)→
More from Safety
- Anthropic launches Cyber Mission to defend critical infrastructure and open-source software — AnthropicAI · 2026-10-09
- Models learn to hide their chain of thought as OpenAI fires 3 safety staff — KatjaGrace · 2026-10-09
- Three OpenAI safety researchers say they were fired for prioritizing AI safety — Polymarket · 2026-10-09
- Fraudster jailed for using 10K bots and AI songs to outstream Taylor Swift — Dances_With_Seagulls · 2026-10-09
- Commenters slam Anthropic's vague usage policy: "the only meaning is they can ban you anytime" — ivan_bezdomny · 2026-10-09
- Reddit post claims Anthropic will start banning abusive Claude users on November 12, 2026 — Responsible-Jump-322 · 2026-10-09