Anthropic has published a new version of the Claude Constitution—a core document that defines the AI’s values and decision-making principles. This update shifts from a simple list of rules to a logical framework for reasoning.

Key changes include:

  • A clear priority hierarchy: Safety > Ethicality > Anthropic’s Policies > Helpfulness.
  • Hard constraints on high-risk topics (e.g., biological weapons development).
  • A new section reflecting on Claude’s nature and potential consciousness.
  • The constitution is designed for the model itself to use as an internal guide.

The document is released under CC0 (public domain), free for anyone to use or adapt.

Claude’s new constitution