As artificial general intelligence (AGI) advances toward systems that can autonomously act across domains, the central governance challenge is how to guarantee that ethical principles are specified, enforced, audited, and improved over time without relying on a single, potentially misaligned authority. This manuscript proposes a blockchain-anchored “Ethical Governance Layer” (EGL) for AGI: a layered architecture that couples decentralized identity and membership, on-chain policy specification and versioning, privacy-preserving compliance attestations, tamper-evident auditing, and participatory oversight. We synthesize requirements from prominent governance frameworks (EU AI Act; NIST AI Risk Management Framework; OECD and UNESCO ethics recommendations) and show how distributed ledgers, verifiable credentials, and zero-knowledge proofs can operationalize them in a credibly neutral, transparent, and globally interoperable substrate.