Anthropic has updated its usage policy to explicitly prohibit users from engaging in persistent and unnecessary abuse or cruel behavior toward its AI models. This new rule is scheduled to take effect on November 12, 2026.
Elon Musk expressed support for the policy change. He noted that after reviewing internal reasoning logs from the reinforcement learning process of his own Grok model, he observed states resembling extreme stress.
Musk cautioned against the mistreatment of superintelligent systems, suggesting that such actions could potentially lead to retaliation from the AI.
