Anthropic Codifies AI Etiquette in Policy Overhaul

Anthropic is updating its usage policy to prohibit "sustained and needless abusive or cruel behavior" toward its Claude AI models, according to reporting from The Register. The new rules, which take effect November 12, target extreme cases where users repeatedly act cruelly without a discernible purpose.

Anthropic clarified that the policy does not apply to model testing, research, dark creative themes, or common user frustration. To enforce these boundaries, the company will rely on a feature introduced in August 2025 as part of its "model welfare" research, which allows Claude to terminate conversations with users who are persistently abusive.

While the etiquette rules draw attention, they are part of a broader policy overhaul. The Register reports that Anthropic has tightened restrictions on deceptive influence campaigns following the use of Claude by commercial organizations, government propaganda offices, and state media to create fabricated news sites and fake accounts. Additionally, the company has clarified that its ban on weapons development extends to components and software, specifically targeting the creation of autonomous vehicles and armed drone guidance systems.

Other key updates include prohibitions on using Claude to develop surveillance tools or to recommend specific individuals for police arrest, investigation, or charging. Anthropic also introduced requirements for users connecting Claude to hardware capable of causing injury, mandating that a qualified human operator be able to monitor the equipment and ensure it remains in a safe state if the AI connection is lost.