AI Trust Glossary · Canonical Definition
Infrastructure
Decentralized Identifier (DID)
A W3C standard for globally unique, cryptographically verifiable identifiers that require no central registration authority. A DID resolves to a DID Document containing verification keys, service endpoints, and metadata about the identified subject.
Explanation
A DID takes the form
did:method:identifier - for example, did:bts:A1B2-C3D4-E5F6-G7H8. The method specifies the rules for creating, reading, updating, and deactivating the identifier. The identifier is method-specific but must be globally unique. Unlike a username or email address, a DID is controlled by its subject - no company, government, or platform can revoke it without the subject's private key.Why it matters for AI agents
AI agents interact with other systems autonomously, often without a human in the loop. They need verifiable identities that prove who they are, what they're authorized to do, and what trust record they carry. Traditional identity systems (OAuth tokens, API keys) are centralized and revocable by the issuing platform. A DID persists on a decentralized registry - when an AI agent presents its DID, any party can verify its identity and trust score without calling a central authority. This is the infrastructure required for EU AI Act compliance at scale.
How Borealis uses it
Every BTS License Key is a proto-DID. The current format
BTS-XXXX-XXXX-XXXX-XXXX maps directly to the did:bts method identifier. Borealis is building the resolution layer silently - first making BTS keys resolve as DIDs, then making BTS trust scores resolve as Verifiable Credentials. The Hedera Consensus Service provides the immutable proof layer required by the W3C DID spec. See the full specification at borealisacademy.com/specs/did-bts/v1.