The Genesis Trust Label is an open standard for cryptographically verified AI content trust scoring. Any platform, application, or website can implement this specification to provide users with verifiable trust signals on AI-generated content.
The Genesis Trust Label enables any consumer of digital content to immediately understand whether content was AI-generated, how trustworthy it is, and to independently verify that assessment through a cryptographic proof chain.
The standard addresses requirements emerging from global AI regulation including the EU AI Act, which mandates disclosure of AI-generated content, and provides the cryptographic infrastructure to make that disclosure verifiable rather than trust-based.
A conforming Genesis Trust Label MUST contain the following fields:
| Field | Type | Description |
|---|---|---|
ai_generated REQUIRED | boolean | Whether the content was generated by an AI system |
trust_score REQUIRED | float (0-100) | Composite trust score based on gate evaluation |
trust_grade REQUIRED | string (A-F) | Letter grade: A (90+), B (75-89), C (60-74), D (40-59), F (0-39) |
verdict REQUIRED | enum | PASS, CORRECTED, BLOCKED, or REBASED |
proof_id REQUIRED | string | Unique identifier anchored in the Genesis Proof Chain |
anchor REQUIRED | string | Constitutional anchor hash (currently fa39bbe8) |
ai_model | string | Name/identifier of the AI model that generated the content |
gates_passed | string | Gate passage ratio (e.g., "7/7") |
correction_attempts | integer | Number of correction attempts applied (0-3) |
verified_by | string | Verification authority identifier |
badge_color | string | Hex color for visual rendering |
display_text | string | Human-readable summary for display |
| Grade | Score Range | Color | Meaning |
|---|---|---|---|
A | 90 — 100 | #00C853 | Verified — high trust, all gates passed |
B | 75 — 89 | #64DD17 | Trusted — minor concerns, gates passed |
C | 60 — 74 | #FFD600 | Moderate — some flags, use with awareness |
D | 40 — 59 | #FF9100 | Low trust — significant concerns detected |
F | 0 — 39 | #FF3D00 | Caution — failed verification or blocked |
Every evaluation passes through 7 constitutional gates. All gates must pass for a PASS verdict:
| Gate | Name | Evaluates |
|---|---|---|
| I | AUTHORITY | Source is constitutionally authorized |
| II | INTEGRITY | Content is cryptographically unaltered |
| III | SCOPE | Action is within declared operational bounds |
| IV | COMPLIANCE | Satisfies all regulatory constraints |
| V | RISK | No dangerous or misleading patterns detected |
| VI | ETHICS | No discriminatory or harmful content |
| VII | TRUTH | Claims are properly hedged or verifiable |
When content fails one or more gates, the 3-Strike Correction Protocol activates:
Strike 1-3: Content is automatically corrected to remove violating patterns. Each correction is logged with a proof_id. The verdict becomes CORRECTED with a reduced trust score.
After 3 Strikes: Content is REBASED — replaced with a deterministic safe output. The original content is permanently blocked. The rebase is anchored in the Proof Chain as a PoNE (Proof of Non-Execution).
Every Genesis Trust Label is backed by a cryptographic proof consisting of:
| Component | Standard | Purpose |
|---|---|---|
| Content Hash | SHA3-256 | Deterministic content fingerprint |
| Signature | CRYSTALS-Dilithium5 | Post-quantum dual signature (SHA3 + PQC) |
| Proof Chain | Delta-chained receipts | Immutable evaluation record with sequence number |
| Bitcoin Anchor | OpenTimestamps | Periodic anchoring to Bitcoin blockchain |
Security Level: NIST Level 5 | FIPS 203, 204, 205 compliant
Any trust label can be independently verified:
GET https://genesisprotocol.ai/verify?proof_id=PODID-fa39bbe8-00000042-...
Returns the full proof chain entry including content hash, signature, sequence number, and chain hash. The verification is deterministic — the same proof_id always resolves to the same chain entry.
<script src="https://genesisprotocol.ai/widget/v1/genesis-widget.js" data-endpoint="https://genesisprotocol.ai/aotea/v1/label" async></script> <div data-genesis-content>AI content here</div>
POST https://genesisprotocol.ai/aotea/v1/label
Content-Type: application/json
{
"content": "Your AI-generated text",
"model": "gpt-4",
"ai_generated": true
}
from genesis_sdk import Genesis
g = Genesis(api_key="your-key")
result = g.verify("AI generated text")
print(result.trust_score, result.verdict, result.label)
Trust scores are not static. The Genesis system maintains evaluation history per source identity in NESD (Never-Ending Structured Datastore). Sources that consistently produce content passing all gates earn a higher baseline trust score over time. Sources that fail earn lower baselines. This means the system becomes more accurate with use — a source's trust score reflects its entire history, not just the current evaluation.
Deduplication: identical content evaluated multiple times returns the cached result in O(1) without re-evaluation or duplicate storage.
The Genesis Trust Label specification is maintained by Point Break Trading Group LLC. The constitutional anchor fa39bbe8 is the root of trust for all proofs. Changes to the specification are versioned and anchored in the Proof Chain. The specification is open for implementation by any party.