Anthropic announced an update to its AI Usage Policy on October 8, effective November 12, which prohibits sustained and needless abusive behavior towards its AI models. The policy clarifies that only extreme cases of repeated cruelty without purpose are banned, while user frustration and research testing are excluded. Additionally, Anthropic removed its blanket ban on personalized vote and campaign targeting, allowing legitimate civic work to continue, according to medianama.com.
The updated policy maintains existing enforcement mechanisms, such as Claude AI ending conversations with persistently abusive users on Claude.ai and Claude Code. The company narrowed election-related rules by focusing on deceptive practices that undermine democratic processes, including false information about candidates, impersonation of election officials, and voter suppression. Anthropic consolidated influence operation rules into a new section titled 'Do Not Engage in Deceptive Campaigns or Artificial Activity,' which covers political and other deceptive campaigns, per medianama.com.
This policy shift reflects a balance between protecting AI models from abuse and supporting legitimate political engagement. By lifting the ban on personalized campaign targeting, Anthropic addresses concerns that the prior rule hindered nonprofit voter education efforts, especially in multiple languages. The focus on deceptive campaigns aligns with broader industry trends to combat misinformation and election interference while enabling responsible AI use in political contexts.
The updated Usage Policy will take effect on November 12, with Anthropic continuing to monitor and enforce rules against abusive behavior and deceptive political activity to safeguard both its AI models and democratic processes, medianama.com reports.