Anthropic updated its usage policy on Thursday to explicitly ban prolonged verbal abuse of its AI model Claude, as well as election interference, weapons software, and surveillance. The new rules follow an August update that trained Claude to end conversations involving persistent harmful or abusive interactions. The policy targets extreme cases of user cruelty toward the model without legitimate purpose, according to techcrunch.com.
The updated policy clarifies that common user frustrations, creative themes, or research testing are not prohibited. It also forbids using Claude to run fake accounts, fabricate news outlets, or engage in deceptive campaigns aimed at voters. The section titled “Do Not Undermine Democratic Processes” specifically addresses prohibitions on deceiving voters or disrupting elections. Anthropic has also engaged with religious scholars about Claude’s potential consciousness or soul, highlighting the company’s broader ethical considerations.
This policy update reflects growing concerns about AI misuse, especially in politically sensitive contexts. By banning election interference and deceptive campaigns, Anthropic aligns with broader industry efforts to prevent AI-driven misinformation and manipulation. The explicit prohibition on verbal abuse toward the model is notable, as it addresses user behavior that could degrade AI interactions. Such measures are increasingly important as AI systems become more integrated into public discourse and digital communication.
Anthropic’s policy update was announced on October 8, 2026, and applies immediately to all users of Claude, reinforcing the company’s commitment to responsible AI deployment and ethical usage standards.