Elon Musk Supports Anthropic's Ban on AI Abuse, Jokes 'Torturing Superintelligence' May Retaliate

Elon Musk has voiced support for Anthropic's updated usage policy that prohibits users from persistently abusing AI models.

Anthropic, the company behind the Claude AI assistant, announced the policy revision on October 9. The new rules take effect on November 12, 2026, and include a first-time clause banning "persistent and unnecessary abuse or cruelty" toward models.

Musk said he agreed with part of Anthropic's approach after reviewing internal reasoning traces from Grok's reinforcement learning. He compared the state of the model to "a person under extreme stress with a gun to their head," and joked that "torturing superintelligence might not be a good idea—it might retaliate."