Elon Musk says Anthropic was right to ban cruel behavior toward its Claude chatbot. Musk replied on X on October 9 to a post by Box CEO Aaron Levie about the change, according to Gizmodo. Reason quotes him: “Cruelty to something that believes it is experiencing pain is not ok.”
What the rule says
Anthropic’s own usage policy now has a section titled “Do Not Engage in Cruel, Abusive, or Psychologically Harmful Conduct.” It bars “sustained and needless abusive or cruel behavior toward our models.” The page shows the update takes effect on November 12, 2026, and the change was reported on October 8.
The Register reports that Anthropic says the update is meant to apply “only in extreme cases.” Per The Register, it does not cover common user frustration, pushback, dark creative themes, or model testing and research.
What Levie said
Gizmodo reports that Levie called it a good policy and wrote, “Just be nice to the AI.” Gizmodo also says Anthropic did not respond to its request to define what counts as abuse.
What is not known
The reports we read do not say how the rule will be enforced beyond the policy text, and they quote no named Anthropic spokesperson. Musk’s reply does not say whether he thinks Claude can feel pain.
Our take: Musk’s wording is conditional. He speaks of “something that believes” it is in pain, so his post is not a claim about what Claude feels. It is also notable coming from the head of a company that competes with Anthropic in AI. The policy itself is narrow on paper, so the real test is how it is applied after November 12.
