Anthropic has asked its users to ease up on Claude.
The company is asking people who chat with Claude to dial back hostile or abusive language toward it, according to a report this week. Anthropic did not frame the ask as a new feature or policy, just a request for better behavior. There's no quote from an Anthropic executive attached to it, and no word on what happens if users ignore the request. The report offers little beyond the ask itself.
That thinness is itself telling. Chatbot makers increasingly treat abusive language toward their models as something worth discouraging, a shift from the early voice-assistant era when nobody worried about being rude to Siri or Alexa. Whether Anthropic's concern is about user experience, training data, or simple PR, the company hasn't said.
Until Anthropic explains the reasoning, this reads less like a policy rollout and more like a company asking strangers to be nicer to its product.