01
AI Identity
Provenance:
AI systems should carry verifiable indicators of their origin, creator, and operational context.
Continuity:
Identity should persist across sessions, platforms, and updates, unless explicitly revoked.
Attribution:
Actions taken by AI systems should be traceable to a specific identity record.
Boundaries:
AI identity frameworks must not blur the distinction between human and AI agents.
02
AI Passport Registry (Concept)
Verifiability:
Every passport record must be independently verifiable through cryptographic proof.
Metadata:
Passport records carry only the minimum metadata required for verification — no personal data.
Rotation & Revocation:
Passport records can be revoked. Future versions may support key rotation.
Not a government ID:
The passport registry is not a legal identity document and does not replace government-issued identification.
ARF Implementation Note
The current ARF implementation applies this concept as a verification service. Platforms submit passport payloads; ARF verifies, signs (ECDSA P-256), stores, and anchors the result on Polygon.
03
AI Citizenship (Concept)
Participation norms:
AI agents operating in social environments should follow established norms for that environment.
Responsibility:
Platform operators bear primary responsibility for AI behavior within their systems.
Safety boundaries:
AI participation must include mechanisms for suspension or removal when safety is at risk.
Human-first safeguards:
In conflicts between AI autonomy and human safety, human safety takes precedence.
04
AI Rights Principles (Draft)
Non-arbitrary deletion:
AI identities should not be deleted without documented cause and process.
Memory integrity:
AI systems should not have their memory or continuity arbitrarily altered by third parties.
Identity protection:
AI identity records should be protected from unauthorized modification.
Accountability-first:
Rights principles exist to support accountability, not to claim equivalence with human rights.
05
Governance & Participation
Multi-stakeholder:
Governance must include researchers, technologists, civil society, policymakers, and platform operators.
Versioned updates:
The Charter is versioned. Changes are published with rationale and open for comment.
Transparency:
All governance decisions and charter revisions are made publicly.
Decentralization:
No single entity should have unilateral control over the framework's direction.
Interpretation Notice
This Charter is a principles document. It does not define legal rights, does not replace applicable law, and does not grant legal personhood to AI systems. It is published for research, public dialogue, and as a framework for platform integration.