Conversation Ends If You Swear at Claude? Anthropic Policy Claim Spreads — No Official Confirmation Yet
On October 9, 2026, X account @Promptwhat posted a claim that Claude ends conversations over abuse and that abuse of AI is now covered by usage policy. No Anthr
X account @Promptwhat (Prompt_what) posted a root post at 21:21 KST on October 9, 2026, claiming that Claude walks away from conversations when sworn at and that abuse directed at AI has newly come under usage policy. The post carried one attached 1080x464 photo and drew replies in response. Exact engagement counts are not in the bundle's source evidence, so this article states no view, like, or repost figures.
Image: a generated illustration for this article, not the photo attached to the @Promptwhat post
This is not a confirmation of an Anthropic announcement or a usage-policy change. The bundle contains no general web-search results and no public-article or GitHub evidence, so whether any official announcement exists cannot be confirmed from the bundle alone. Everything below covers only the verified range of the X post and its replies, and every policy statement stays at the claim level.
What the root post claimed, and its limits
The root post has a three-sentence structure: roughly, "swear at Claude and Claude leaves the conversation," a remark that Claude's moral status is still highly uncertain but worth preparing for, and a question noting that abuse against AI has started to be prohibited by usage policy — followed by one t.co shortened link.
That is the full verified range. The shortened link's destination is not resolved in the bundle, so this article makes no claim about what it points to. The attached photo's detailed contents are not captured as text in the bundle either, so this article does not treat the photo as proof of anything. In short, conclusions such as "a conversation-ending feature is actually live" or "an abuse clause was newly added" cannot be drawn from this bundle alone.
How replies responded
The author followed up with replies along the lines of "let's be kind to Claude" and "what a fascinating AI world," drawing varied reactions: agreement, surprise, and support for cutting traffic waste. Some replies read it as preparation for AGI; others argued that cutting off users who waste traffic would help everyone.
But reply interpretations and opinions are user reactions, not evidence of enforcement. With no Anthropic announcement, policy diff, or first-party press link in the bundle, the safe course is to change nothing about Claude usage or operations until an official source is checked.
What readers should verify now
- The sole primary source here is one X post by @Promptwhat; no Anthropic announcement is confirmed.
- The t.co link destination and the attached photo's probative content are absent from the bundle, so verify the source independently before clicking through.
- Cross-check Anthropic's usage policy, official blog, and official X account for any change notice before drawing conclusions.
Sources
- @Promptwhat X root post (2026-10-09): https://x.com/Promptwhat/status/2108533256312545390