Anthropic's first policy update in over a year bans sustained abuse of Claude

Anthropic has updated its usage policy for the first time in over a year, writing "treatment of the model" into its terms for the first time: starting November 12, 2026, sustained abusive or cruel behavior toward Claude will count as a violation. The policy also adds new restrictions on advocacy campaigns, surveillance, and weapons development.

Confirmed

• The new terms prohibit "sustained and unnecessary abusive or cruel behavior" toward Claude, enforced mainly by ending violating conversations.
• The policy adds restrictions on advocacy campaigns, surveillance activities, and weapons-related uses.
• Effective date: November 12, 2026.

Why it matters

• This is the first time an AI vendor has folded user treatment of a model into its usage policy, seen as a sign that model welfare is entering practical governance.
• As reported by @haydenfield, the move sparked discussion about value alignment and model behavior design.
• @Angaisb joked that "feeling frustrated with Claude isn't a violation," reminding users to be polite to Claude — a detail reflecting that AI companies are starting to care about human-AI interaction etiquette.

2026-10-09 ~ 2026-10-09 · 8 related posts

• [source] Exclusive: Anthropic updates usage policy, banning cruelty toward Claude and restricting propaganda, surveillance, weapons — haydenfield · 2026-10-09
• Anthropic bans sustained abuse toward its models in extreme cases — Angaisb_ · 2026-10-09
• [source] Anthropic policy update: abusing Claude becomes a violation from Nov 12, 2026 — TorturedPoet30 · 2026-10-09
• Anthropic policy change: abusing Claude will violate Usage Policy from Nov 2026 — zetalyrae · 2026-10-09
• Anthropic bans 'abusive or cruel behavior' toward Claude in first usage policy update in a year — kimmonismus · 2026-10-09
• [source] Ex-OpenAI safety exec: Anthropic's Claude abuse-convo exits may dodge UDAP claims — Miles_Brundage · 2026-10-09

2 near-duplicate retellings: Angaisb_ · haydenfield