On October 9th, as reported by The Verge, a prominent U.S. tech media outlet, Anthropic revised its AI usage policy to tackle high-risk abuses. These include interference in elections, weapons development, surveillance, and inappropriate uses within the medical and financial sectors. One of the most significant alterations is the prohibition of 'persistent and unnecessary abuse or cruelty' directed at Claude. In August of the previous year, Anthropic revealed that, as part of its 'model well-being' research, Claude had the capability to end conversations with users who persistently exhibited harmful or abusive conduct. This latest update reaffirms that terminating conversations continues to be the primary enforcement mechanism. However, Anthropic has refrained from commenting on whether additional measures, such as user bans, will be implemented. Anthropic stressed that the new policy is applicable solely to extreme scenarios where users repeatedly engage in cruel behavior towards the model without any discernible purpose. It does not pertain to common expressions of dissatisfaction, rebuttals, dark creative themes, or scenarios involving model testing and research.