Anthropic can now stop users who are cruel to Claude

AI company Anthropic, the maker of the Claude chatbot, has changed its usage policy. The rules now prohibit "persistent and unnecessary" abusive or cruel behavior toward the company's AI models. That is according to several tech outlets, including TechSpot and Android Authority.

As some headlines have put it, you can be banned for being too unkind to an AI. That is a simplification, however.

The conversation is ended first

According to the reports, the primary tool is that Claude can end the conversation itself with a user who is repeatedly cruel. According to the coverage, Anthropic has not said whether the company will also use other sanctions, such as closing the user's account entirely.

Extreme cases only

The rule is meant for extreme cases in which a user is cruel over and over again for no purpose. It does not apply to ordinary frustration when the AI answers incorrectly. Nor does it apply to pushing back against the AI, writing dark themes in fiction, or testing and researching the models.

So if you have ever written something sharp to a chatbot that did not understand you, you are not in the danger zone.

Why is this interesting?

AI models like Claude are computer programs, and there is no consensus that they can feel anything. Even so, the rule treats them somewhat as something that should be protected from mistreatment. The outlet Bullsource writes that the change puts to the test the question of whether AI can have a form of personality.

For ordinary users, the consequences are small today. But it shows that AI companies are beginning to write rules about how we talk to the machines, and not only about what we use them for.

I have not been able to read Anthropic's own text directly, so the exact wording and details about enforcement should be checked against the company's official usage policy.