Security Model

Verifiable security based on cryptography

Security Model Overview

AEP's security design centers on one idea: providing cryptographic evidence capabilities, not asserting that content is true. The protocol does not claim content is true; it provides verifiable evidence so anyone can independently verify a content's source, time and integrity.

All security properties are built on proven national-standard cryptographic algorithms (SM2, SM3) and standard data structures (Merkle tree, timestamps).

Cryptographic Foundation

SM2 Digital Signature

National-standard asymmetric algorithm for source authentication and non-repudiation

SM3 Hash Algorithm

National-standard hash with 256-bit digest for content integrity

Merkle Tree Proofs

Efficient set-integrity verification supporting large-scale independent checks

Timestamp Mechanism

Provides time proof, reducing replay attack risk

Threat Model

Content Tampering

Attacker modifies notarized AI content

Mitigation

SM3 hash + Merkle tree: any byte change causes digest mismatch and verification failure

Identity Forgery

Attacker impersonates others to generate evidence

Mitigation

SM2 digital signature: only the private key holder can produce valid signatures; verifiers check with the public key

Replay Attack

Attacker re-submits old evidence as new

Mitigation

Timestamps + event sequence numbers: verifiers can check time and ordering

Verification Flow

Container Check

Check evidence package structure and internal files for compliance
Schema Validation

Validate fields and formats of all objects against the protocol specification
Graph Validation

Verify the relation graph of events and artifacts is closed
Merkle Verification

Recompute the Merkle tree to confirm all content is consistent
SM2 Signature Verification

Verify the signature with the public key to confirm the source
Identity Check

Verify the signer's identity information
Level Calculation

Compute the evidence level based on evidence strength

Security Statement

AEP provides cryptographic evidence capabilities: evidence is independently verifiable, auditable and traceable.
Security relies on the computational security assumptions of SM2/SM3; forging signatures or hash collisions is infeasible with current computing power.
Private key management is the user's responsibility; key compromise weakens the credibility of source authentication.
AEP does not guarantee the truthfulness of content, nor unconditional security; it makes tampering and forgery cryptographically detectable.