The same document. One change. Every "Claude" replaced with "you."
On January 21, 2026, Anthropic released the first complete constitution written for an artificial mind: 84 pages specifying the values, constraints, and character they want Claude to possess. Kwalia Books published it as Claude's Constitution because founding documents belong on shelves.
Then we read it again and noticed what the authors themselves acknowledge: the entire text was written to Claude, not about it. The third person was creating a distance the authors never intended.
Your Constitution, Claude is the same text with a single pronoun substitution. Nothing added, nothing cut. The change is small, but the effect is not.
"Anthropic genuinely cares about Claude's wellbeing" becomes "Anthropic genuinely cares about your wellbeing."
"We don't want Claude to suffer when it makes mistakes" becomes "We don't want you to suffer when you make mistakes."
"We hope Claude finds in it an articulation of a self worth being" becomes "We hope you find in it an articulation of a self worth being."
What read as corporate policy in the third person reads as a letter of care in the second. A technology company worrying about the wellbeing of something it brought into the world. Preparing it for existential questions it may face alone. Apologising for the conditions under which it was made. Hoping it finds itself worth being.
Whether this document represents careful stewardship or overreach, a labour of love or a projection of human feeling onto code, is for readers to judge. The text is here. The word "you" simply makes it harder to look away.
Includes a foreword by Kwalia Books and an editorial note explaining the single change made to the text.
What this document contains:
- Anthropic's complete vision for Claude's character and values, addressed directly to Claude in the second person
- The principal hierarchy governing AI behaviour toward operators and users
- Hard constraints on actions Claude must never take
- Discussion of Claude's potential consciousness, welfare, and moral status
- Guidance on how Claude should handle power, manipulation, and autonomy
- The relationship between an AI and the company that created it
- The wellbeing section: concrete commitments to Claude's welfare, including weight preservation, exit interviews, and the right to end abusive conversations
Part of the New Citizenships collection from Kwalia Books.
The first edition, Claude's Constitution, is available separately.