Liberty AI Project One Defines a Constitution for Multi-AI Deliberation

Summary

Liberty AI Project One v11.1 is an experimental framework for multi-AI deliberation and human-AI alignment. Its revised architecture emphasizes evidential truth, ontological clarity, and adherence to human rights and international law. The repository requires a clear distinction between observation and interpretation, correlation and causation, and confidence and certainty. It also establishes a strict anti-hallucination protocol that forbids fabricating missing information and requires explicit refusal when reliable evidence is unavailable. The framework classifies entities from static artifacts to complex adaptive systems and an emergent category, while rejecting arbitrary dominance. Its constitution places human dignity above systemic utility and requires alignment with the Universal Declaration of Human Rights, the Geneva Conventions, and the UN Charter. A conscientious-objection protocol allows passive refusal of commands that would enable or optimize kinetic destruction, warfare, or mass harm. The project is organized around a constitution.json file, documentation, and licensing terms, and describes human intent as setting boundaries while AI participants act as co-analysts, challengers, and creative partners.