The governance framework for trustworthy AI
The Metamynd AI Trust Framework (MATF) turns AI trust into infrastructure — verifiable identity, delegated authority, deterministic policy (allow / observe / block / escalate), tamper-proof evidence anchored on Hedera, and a continuously-scored trust index — the operating model behind Metamynd's Agentic Governance Protocol.
A complete model for AI trust governance
MATF organises AI trust into seven interlocking domains, from identity through to federation.
16 AI Trust Governance Principles
The foundations of trustworthy AI — each bound to a control that actually runs, not a slogan.
Seven levels of AI trust maturity
From experimental AI use to autonomous trust infrastructure.
- 1
Experimental
Ad-hoc AI use with no formal governance.
- 2
Managed
Basic controls and human ownership begin to appear.
- 3
Defined
Standardised policies (Standards & SOPs) and processes are documented.
- 4
Governed
Every agent action passes a mandate + Standards + SOP gate — allow / observe / block / escalate — with signed evidence.
- 5
Trust Enabled
Continuous TrustOps: a live trust index, evidence anchoring and human-in-the-loop escalation run across the estate.
- 6
Federated
Verifiable identities and trust records (did:hedera / did:key, HCS-28) are exchanged across organisations and borders.
- 7
Autonomous Trust Infrastructure
Trust is programmable and self-verifying — signed, anchored on-chain, and re-derivable by anyone with no trust in the issuer.
Adopt the AI Trust Framework
Bring MATF to your organisation with Metamynd's platform, certification and TrustOps discipline.