PRINCIPLES & RECOMMENDATIONS
AI governance principles shape laws, standards, organizational policies, assurance systems, and public expectations. The library explains what each source says, who issued it, what force it carries, and how its expectations can be translated into evidence-bound TA-14 governance conditions.
WHAT THIS CATEGORY CONTAINS
Principles and recommendations can define values, objectives, safeguards, expected conduct, or public-interest boundaries. They may influence legislation, standards, contracts, procurement, internal policy, or assurance. The library does not mislabel them as binding law, certification, or proof of implementation.
SOURCE CATALOG
Each module preserves the source authority, publication history, intended audience, normative force, themes, relationships, implementation questions, evidence expectations, and TA-14 route pathways.
Organisation for Economic Co-operation and Development
Intergovernmental principles
High-level principles for trustworthy AI, including inclusive growth, human-centered values, transparency, robustness, safety, accountability, and responsible stewardship.
Open source module→United Nations Educational, Scientific and Cultural Organization
International recommendation
A rights-based and ethics-focused governance instrument addressing human dignity, fairness, environmental well-being, diversity, oversight, accountability, and public-interest safeguards.
Open source module→International, regional, and domestic authorities
Rights-based governance source
A cross-cutting source category connecting AI systems to dignity, privacy, equality, due process, freedom of expression, non-discrimination, remedy, and procedural fairness.
Open source module→Governments, institutions, civil society, and sector bodies
Policy and advisory principles
Principles intended to protect public welfare, institutional legitimacy, democratic accountability, accessibility, safety, and responsible deployment.
Open source module→FROM PRINCIPLE TO GOVERNED CONDITION
A principle such as accountability, fairness, transparency, or human oversight cannot govern execution until its scope, owner, evidence, threshold, decision effect, review path, and outcome record are made explicit.
INTERPRETATION BOUNDARIES
A principle may influence law without itself being binding law.
A recommendation may guide institutions without proving implementation.
Ethical language does not automatically establish operational controls.
Alignment claims require evidence, ownership, and preserved review.
Conflicts between principles must remain visible and unresolved when necessary.
TA-14 routes preserve how a principle became a bounded execution condition.
BOUNDARY
A statement that an AI system is ethical, responsible, transparent, fair, human-centered, or rights-respecting remains a claim until the supporting authority, interpretation, evidence, control, decision, execution, and outcome are preserved and reviewable.
Live cumulative activity recorded across the public Exchange surface.