Anthropic Updates Usage Policy to Ban Sustained Cruelty Toward Its AI Chatbot Claude

• •Anthropic's revised usage policy bans sustained and purposeless abusive or cruel behavior toward Claude, the first such restriction from a major AI company.
• •The company has not specified what constitutes abusive content, though routine frustration and criticism of Claude remain permitted.
• •Broader policy revisions also cover election interference, deceptive campaigns, weapons development, surveillance, and high-risk health and financial uses.
• •Recent scientific research has explored inducing emotional states in AI chatbots, finding that manipulating and then stabilizing LLM responses can yield useful outcomes.
• •The policy update is Anthropic's first in over a year, reflecting new categories of misuse the company identified during that period.

Sources

Related Stories