Elon Musk backs ban on abusive AI treatment as Anthropic tightens policies

On October 13, 2026, global tech magnate Elon Musk endorsed Anthropic's updated policies prohibiting users from engaging in persistent and purposeless abusive behavior toward AI models. The move follows Anthropic's recent revision of its Claude usage policy, which will take effect on November 12, 2026. The policy now explicitly bans 'repeated and unnecessary abuse or cruel treatment' of AI systems, targeting users who repeatedly and without clear intent subject models to harmful actions. Musk expressed support for Anthropic's stance, drawing a parallel between the treatment of AI and a person under extreme duress. He remarked that 'tormenting superintelligence (SI) might not be a good idea. It could retaliate.' His comments came after reviewing internal reinforcement learning trajectories of Grok, which he described as akin to being 'pointed a gun at the head.'