Anthropic updates usage policy, banning extreme abuse of Claude and widening misuse rules

Mixed · 2Anthropic updated its usage policy on Thursday 8 October 2026, its first change in over a year, to address new and high-risk cases of misuse including election interference, weapons development, surveillance, and health and financial uses. 1 2 Mixed · 2The most attention-getting change prohibits "sustained and needless abusive or cruel behavior" toward Claude. 1 2 Mixed · 2Anthropic said the rule "is meant to apply only in extreme cases, where users repeatedly act cruelly toward our models, with no discernible purpose" and does not apply to "common versions of user frustration, pushback, dark creative themes, or model testing and research." 1 2

Mixed · 2Since an August update, Claude has been able to end conversations with persistently harmful or abusive users, as part of Anthropic's research into "model welfare"; the new policy keeps ending conversations as the "primary enforcement mechanism", and Anthropic did not tell The Verge whether other measures such as user bans are planned. 1 2 Mixed · 2Scattered existing restrictions were gathered into a ban on deceptive commercial or political campaigns, covering efforts to obscure who is behind a message or amplify content through fake accounts or posts, and the election section bars voter deception and election disruption, such as misinformation about candidates or how to vote, impersonating officials, or suppressing turnout; TechCrunch notes the section is titled "Do Not Undermine Democratic Processes". 1 2 Left · 1The weapons ban, long public, now extends to software and components that make weapons work and actions like arming drones and other autonomous vehicles, because Anthropic said it had seen multiple attempts to use Claude for weapons guidance and control software. 1

Left · 1Surveillance rules were clarified after Anthropic's last threat intelligence report suggested growing use of Claude-powered surveillance to track political dissidents: tracking people without consent is prohibited, Claude cannot be used to decide who to investigate, arrest or charge, and it cannot be used to build or improve surveillance tools. 1 Left · 1These rules may be modified for contracts with certain governmental customers if Anthropic judges the contractual restrictions and safeguards adequate; The Verge notes the company has previously contracted with the US military. 1 Left · 1A new rule also requires that when Claude is connected to hardware that takes autonomous physical actions and could cause injury, a qualified operator must be able to observe the equipment and stop it if needed; The Verge says it is unclear whether that operator must be human. 1

Mixed · 2The Verge placed the model-welfare change in the context of Anthropic's openness to the possibility that its models are conscious, citing CEO Dario Amodei's remark that "We don’t know if the models are conscious", and contrasted it with Microsoft AI's stance against model welfare; TechCrunch linked it to Anthropic's outreach to religious scholars, in some cases seeking to convince them Claude could be conscious or have a soul. 1 2