Whitepaper · v2.1

Human Authorship
Protocol

A technical framework for defining, measuring, and verifying human creative contribution within AI-assisted works - including legal considerations, dispute resolution, blockchain architecture, and C2PA integration.

Published: 2026

Version: 2.1

Status: Open Standard

Author: Aquarius Maximus

Abstract

The Human Authorship Protocol (HAP) is a framework designed to define, measure, and verify human creative contribution within AI-assisted works. As generative AI systems increasingly produce content with minimal human input, traditional definitions of authorship and ownership are being challenged. HAP introduces a structured approach to capturing the human role in creative systems, enabling attribution, ownership, and monetization in an AI-driven economy.

Introduction

Artificial intelligence has drastically reduced the barrier to content creation. Music, art, writing, and media can now be generated with minimal effort. However, this introduces a fundamental problem: the separation of creation from authorship. Current systems fail to adequately represent human intent, creative direction, and iterative decision-making. HAP addresses this gap by establishing a new framework that centers the human creative process.

The Authorship Problem

2.1 Output vs Process

Traditional authorship focuses on outputs: a song, a painting, a written work. AI systems disrupt this by generating outputs without clear human authorship.

2.2 The Missing Layer

The missing layer is the system of decisions that led to the output. Without capturing this layer, authorship becomes invisible and unverifiable.

Defining Human Authorship

“The structured set of human decisions that guide and shape AI-generated outputs.”

This includes prompt systems (recipes), creative inputs, selection and refinement, curation, and identity-based contributions.

Core Components

4.1 Recipes (System Layer)

Recipes are structured frameworks that define style, structure, tone, and output constraints. Recipes are repeatable and represent system-level authorship. They are ownable, licensable, and sellable independently of any output they produce.

4.2 Voice (Identity Layer)

Voice represents human vocal input and identity-based contribution. Voice is treated as a distinct asset class with granular licensing permissions. Voice data is biometric in nature and subject to strict privacy protections (see Section 10).

4.3 Inputs (Creative Layer)

Inputs include lyrics, composition guidance, edits and modifications, and any explicit creative direction provided by the human.

4.4 Iteration (Process Layer)

HAP captures version history, decision points, and output selection. Rejection decisions are as meaningful as selection decisions - they document creative judgment.

4.5 Curation (Selection Layer)

The final act of selecting which output to keep is a distinct human contribution. Curation is tracked as its own component in the Human Contribution Score.

Human Contribution Score (HCS)

“HCS tiers are descriptive indicators only. Copyright eligibility is determined by applicable law and cannot be established by any score or framework.”

HAP introduces a quantitative model: the Human Contribution Score (HCS).

5.1 What the HCS Measures

The HCS is the sum of all human-driven creative components. AI's role is the remainder - it is not listed as a component of the human score. HCS = Sum of all human-driven components AI Remainder = 1 − HCS Important: HCS is a documentation metric, not a legal threshold. U.S. courts and the Copyright Office focus on qualitative human authorship, not a quantitative score. A high HCS does not guarantee copyright protection; a lower HCS with deep qualitative creative shaping may support a stronger claim than a high HCS built on minimal decisions. No score produced by HAP constitutes legal advice or a determination of copyright eligibility.

5.2 Weighting Rationale

Weights are assigned based on three criteria: Intentionality (does this component require deliberate human decision-making?), Reproducibility (would removing this component significantly change the output?), and Defensibility (can this component be documented as evidence of authorship?). Note on Curation: Courts have been inconsistent on whether selection alone constitutes authorship. Feist v. Rural Telephone (1991) held that selection can be creative, but only where it reflects genuine intellectual judgment. HAP tracks curation as a documented decision - it does not imply that curation alone creates a copyright claim.

Component	Intentionality	Reproducibility	Defensibility	Default Weight (Music)
Recipe	High	High	High	30%
Inputs	High	High	High	25%
Voice	High	High	Medium	20%
Iteration	Medium	Medium	High	15%
Curation	Medium	Low	Medium	10%

05b

Domain Profiles & HCS Tiers

Different creative industries have different contribution patterns. HAP defines three default domain profiles:

Music (Default)

Recipe 30% / Inputs 25% / Voice 20% / Iteration 15% / Curation 10%

Visual Art

Recipe 35% / Iteration 25% / Inputs 20% / Curation 15% / Voice 5%

Written Content

Inputs 40% / Recipe 25% / Iteration 20% / Curation 10% / Voice 5%

HCS Range	Tier	Meaning
0.80 – 1.00	Primary Human Authorship	Human creative direction dominates throughout
0.50 – 0.79	Collaborative Authorship	Significant human input with substantial AI generation
0.20 – 0.49	AI-Assisted Creation	AI-led output with human curation and light direction
0.00 – 0.19	AI-Generated	Minimal human contribution beyond prompt input

Data Structure

Each creative work is associated with a structured authorship record:

authorship-record-v2.1.json

{
  "hap_version": "2.1",
  "record_id": "hap_record_20250101_001",
  "domain_profile": "music",
  "recipe": {
    "recipe_id": "hap_recipe_v2_001",
    "recipe_cid": "bafybeig...abc123",
    "storage_provider": "pinata",
    "description": "Lo-fi hip hop, 70-85 BPM, minor key, vinyl texture"
  },
  "voice": {
    "voice_id": "hap_voice_creator_001",
    "voice_model_cid": "bafybeig...def456",
    "storage_provider": "pinata",
    "voice_permissions": {
      "licensed_uses": ["music_generation", "remix"],
      "prohibited_uses": ["advertising", "political_content"],
      "expiry": "2027-01-01T00:00:00Z",
      "revocable": true
    }
  },
  "inputs": [{ "input_id": "lyric_v1", "input_cid": "bafybeig...ghi789", "storage_provider": "pinata" }],
  "iterations": [
    { "version": "v1", "selected": false, "notes": "Tempo too fast." },
    { "version": "v2_selected", "selected": true, "notes": "This one. Keep." }
  ],
  "hcs_components": {
    "recipe": 0.30, "inputs": 0.25, "voice": 0.20,
    "iteration": 0.15, "curation": 0.10
  },
  "contribution_score": 0.90,
  "hcs_tier": "primary_human_authorship",
  "onchain_anchor": {
    "chain": "polygon",
    "transaction_hash": "0xabc123...",
    "anchored_hash": "sha256:9f86d081...",
    "timestamp": "2025-01-01T00:00:00Z"
  },
  "owner_identity": "0x7f3a...b291"
}

Blockchain Integration

HAP supports optional on-chain anchoring using a hybrid architecture. Compact cryptographic proofs go on-chain; actual creative assets go off-chain in a content-addressed system (IPFS).

Stored On-Chain

Hash of authorship data · Ownership reference · Timestamp · Creator wallet address

Stored Off-Chain

Audio files · Full prompt structures · Voice model data · Large creative datasets

The Anchoring Process

1. Finalize the record - all off-chain assets uploaded to IPFS, CIDs written into JSON. 2. Hash the record - SHA-256 hash of canonical JSON computed. 3. Write to chain - hash, creator wallet, and Unix timestamp written to smart contract. 4. Store the record - complete JSON record uploaded to IPFS.

Chain	Pros	Cons	Recommended Use
Guapcoin (EVM)	Native HAP chain, low cost, EVM-compatible	Purpose-built for HAP	HAP authorship records - recommended
Ethereum Mainnet	Highest security, widest trust	High gas costs	High-value commercial works
Polygon	Low cost, EVM-compatible	Less decentralized	High-volume platforms
Arweave	Permanent storage built-in	Different model, not EVM	Archival / long-term records
Base	Low cost, Coinbase-backed, EVM	Newer, less battle-tested	Consumer apps, music platforms

Ownership Model

8.1 Output Ownership

The generated work - the final audio, image, or written piece produced by the AI system.

8.2 System Ownership

The recipe and creative framework - the human-designed process that guided the output. Recipes are ownable and licensable independently. Legal basis note: In U.S. IP law, systems, methods, and processes are generally not protectable under copyright - that is patent or trade secret territory. Recipe ownership under HAP is structured as a combination of: (1) trade secret protection under the Defend Trade Secrets Act (DTSA) for undisclosed recipes, (2) contractual license enforced through HAP platform terms of service, and (3) sui generis rights defined by the HAP registry. Creators licensing recipes commercially should obtain qualified IP counsel to determine the appropriate legal structure in their jurisdiction.

8.3 Identity Ownership

Voice and creator identity - the human's unique biometric contribution that can be licensed with granular permissions and revoked at any time.

Monetization Model

“Securities law notice: Fractional ownership interests in creative works sold to multiple parties on decentralized platforms may constitute investment contracts subject to SEC regulation under the Howey test, particularly where buyers expect returns based on others' efforts. Platform implementers offering contribution-based revenue sharing or fractionalized creative assets must obtain qualified securities counsel before deployment.”

HAP enables new economic models for creative work. The terms below have distinct legal meanings: license conveys permission to use under defined terms without transferring ownership; sale transfers ownership of a specific asset; revenue sharing distributes proceeds proportionally without creating ownership interests.

Recipe licensing - grant permission to use a creative system under defined terms (scope, duration, exclusivity)
Voice licensing - monetize vocal identity under defined permissions (use case, expiry, jurisdiction)
Contribution-based revenue sharing - revenue split proportional to HCS components
Creative asset licensing and transfer on decentralized platforms

Legal Framework

10.1 Copyright and AI-Generated Works

The US Copyright Office has established that copyright requires human authorship (see Thaler v. Perlmutter, 2023). HAP's HCS and its components are designed to document the kind of human creative control that copyright frameworks look for. Works in the Primary Human Authorship tier (0.80–1.00) are most likely to support a copyright claim. Works in the Collaborative Authorship tier (0.50–0.79) may support partial protection. Works below 0.50 are unlikely to qualify under current US standards. Critical caveat: HCS tiers are descriptive indicators, not legal determinations. Copyright eligibility is a qualitative legal question that no score or framework can resolve. A high HCS does not guarantee copyright protection. HAP is a documentation framework, not a legal service.

10.2 Jurisdiction Considerations

United States: The Copyright Office requires human authorship. HAP records document intentional creative decisions directly relevant to this standard. European Union: EU copyright requires a work to be the "author's own intellectual creation." HAP's Recipe and Input components are designed to document this kind of contribution. However, EU member states implement this standard differently - France, Germany, and others have distinct moral rights frameworks and judicial interpretations. EU alignment requires jurisdiction-specific legal review and cannot be assumed from HAP documentation alone. United Kingdom: The CDPA s.9(3) provision grants copyright to the person who made the "arrangements necessary" for the creation. HAP's system-level authorship model is relevant to this framing. However, this provision is controversial and under active review by the UK Intellectual Property Office - it may be narrowed or repealed. Creators should not rely on s.9(3) as a stable legal basis without current qualified UK IP counsel.

10.3 Voice Data and Biometric Privacy

GDPR (EU/EEA): Biometric data is a "special category" under Article 9. Processing requires explicit consent and a DPIA. Important: Regulators have taken the position that pseudonymized or hashed biometric data may still constitute personal data if re-identification is possible in context. The on-chain hash alone does not eliminate GDPR obligations - it is a design mitigation, not a compliance solution. BIPA (Illinois, USA): Requires written informed consent before collecting voiceprints, plus a publicly available retention and destruction policy before collection begins. Statutory damages: $1,000–$5,000 per violation. Critically, courts have interpreted BIPA liability as accruing per individual collection event - aggregate exposure can be substantial. Platforms should obtain qualified BIPA counsel before implementing voice data collection. State law landscape: Biometric and voice-adjacent privacy laws vary widely and are rapidly evolving. Texas (CUBI), Washington, New York, California (CCPA/CPRA), and others have overlapping frameworks. This whitepaper does not constitute a complete survey of applicable law. HAP's design response: Voice models stored off-chain and encrypted. On-chain record contains only a non-reversible hash. Platforms must implement a right-to-deletion workflow.

record.json

"voice_permissions": {
  "licensed_uses": ["music_generation", "remix"],
  "prohibited_uses": ["advertising", "political_content"],
  "expiry": "2026-01-01T00:00:00Z",
  "revocable": true,
  "jurisdiction_restrictions": ["US", "EU"]
}

10.4 HAP Records as Legal Evidence

What makes a HAP record legally useful: Timestamp integrity (on-chain anchoring proves creation date), Chain of custody (iteration history documents the process), Identity binding (owner wallet links record to a specific entity). What HAP records cannot do: They cannot prove that a human (rather than a script) made the documented decisions. They do not constitute legal proof of copyright ownership - they constitute evidence in support of a claim.

10.5 Liability Boundaries

HAP as a protocol does not assume liability for the legality of generated content, the accuracy of HCS scores as legal determinations, or the enforceability of voice licensing terms. Platforms implementing HAP are solely responsible for legal compliance in their operating jurisdictions.

Off-Chain Data Integrity

HAP mandates content-addressed storage for all off-chain assets. The recommended implementation is IPFS, which assigns every file a CID (Content Identifier) - a cryptographic fingerprint. If a file changes by even a single byte, its CID changes, making tampering immediately detectable. Platforms must use an IPFS pinning service (e.g. Pinata, web3.storage) to guarantee off-chain asset persistence. Unpinned assets may be garbage collected and permanently lost.

11.1 Deletion and Privacy Rights

On-chain data (hash, wallet, timestamp) is never personal data under GDPR. Off-chain data can be deleted - the on-chain hash becomes an orphaned proof that something existed without revealing the content. This satisfies GDPR without requiring blockchain modification. Voice model deletion must be a one-click workflow that unpins IPFS data, revokes the CID in the platform index, and logs the deletion event.

Dispute Resolution

“On-chain anchoring creates evidentiary support for priority of creation - it does not confer automatic legal priority. Courts do not automatically defer to blockchain timestamps. Chain-of-title disputes, fraud, or errors in registration may override an earlier anchor. In copyright law, priority is determined by the act of creation, not registration.”

HAP defines a three-tier dispute resolution process for four dispute types: Ownership, Derivation, Voice Identity, and Contribution disputes.

Tier 1 - Automated Similarity Check (Platform Level)

Before registration, the platform runs an automated similarity check. Recipes scoring above 85% similarity trigger a notification to both parties. Voice models above threshold trigger an identity verification request.

Tier 2 - Structured Mediation (HAP Registry Level)

A formal Dispute Notice is submitted. The registered owner has 14 days to respond. A HAP Resolver reviews evidence and issues a recommendation within 21 days. Outcomes: Uphold, Transfer, Co-register, or Escalate. All proceedings logged on-chain.

Tier 3 - External Arbitration

For unresolved or high-value disputes. HAP provides structured evidence packages to recognized arbitration bodies including WIPO (IP-specialized, recommended for cross-border IP disputes), AAA (American Arbitration Association), JAMS, and ICC (International Chamber of Commerce). Note: Smart contracts are not uniformly enforceable as contracts in all jurisdictions. Some jurisdictions require specific formalities for IP license agreements. Parties should confirm arbitration clause enforceability with local counsel before relying on HAP dispute resolution.

Anti-Gaming Protections

Bad faith registration penalty: Voided registration and flagged account. Dispute filing stake: A nominal token deposit required, returned if upheld. Cooling-off period: 90-day bar on re-filing the same dispute without new evidence. Timestamps cannot be backdated: On-chain anchor is set at block confirmation.

12.5 Dispute Record Schema

All disputes generate a structured on-chain record:

record.json

{
  "dispute_id": "hap_dispute_20250201_001",
  "disputed_record_id": "hap_record_20250101_001",
  "dispute_type": "ownership",
  "filed_by": "0xabc1...1234",
  "filed_against": "0x7f3a...b291",
  "filed_at": "2025-02-01T00:00:00Z",
  "status": "resolved",
  "tier": 2,
  "evidence_submitted": [
    "bafybeig...evidence001",
    "bafybeig...evidence002"
  ],
  "resolver_id": "hap_resolver_007",
  "recommendation": "uphold",
  "resolved_at": "2025-02-19T00:00:00Z",
  "onchain_anchor": {
    "chain": "polygon",
    "transaction_hash": "0xdisp789...",
    "timestamp": "2025-02-19T00:00:00Z"
  }
}

12.6 Collaboration and Split Attribution

Multi-party works use a Collaboration Record. HCS components are attributed across multiple owner_identity addresses. Contribution shares must sum to 1.0. All collaborative works should have a signed attribution agreement stored as an off-chain document with its CID in the record.

record.json

{
  "hap_version": "2.1",
  "record_id": "hap_collab_20250101_001",
  "collaboration": true,
  "contributors": [
    {
      "identity": "0x7f3a...b291",
      "role": "recipe_author",
      "hcs_components": { "recipe": 0.30, "iteration": 0.08 },
      "contribution_share": 0.55
    },
    {
      "identity": "0xabc1...1234",
      "role": "vocalist",
      "hcs_components": { "voice": 0.20, "inputs": 0.15, "curation": 0.10 },
      "contribution_share": 0.45
    }
  ],
  "total_hcs": 0.83,
  "hcs_tier": "collaborative_authorship",
  "attribution_agreement_cid": "bafybeig...agreement001"
}

Relationship to C2PA

“C2PA is the provenance layer - it tells you where a file came from. HAP is the authorship and ownership layer - it tells you who made the creative decisions and who owns the rights. C2PA is a shipping manifest. HAP is a deed of ownership.”

The C2PA (Coalition for Content Provenance and Authenticity) specification is an open technical standard for embedding content provenance metadata into media files. It is backed by Adobe, Microsoft, Google, Sony, and others. C2PA and HAP are complementary, not competing.

Dimension	C2PA	HAP
Core question	Where did this content come from?	Who shaped this content, and how?
Primary mechanism	Signed manifest embedded in file	Structured authorship record, on-chain anchored
Human contribution	Not measured	Measured via HCS
Creative system capture	Not captured	Recipe layer
Voice identity	Not addressed	Voice layer with licensing
Ownership model	Not addressed	Output, system, and identity ownership
Monetization	Not addressed	Recipe marketplaces, voice licensing
Blockchain	Not used	Optional on-chain anchoring

c2pa-integration.json

"c2pa": {
  "manifest_identifier": "urn:c2pa:manifest:abc123xyz",
  "claim_generator": "Adobe Firefly 3.0",
  "manifest_cid": "bafybeig...c2pa456",
  "verified": true
}

Worked Example - AI-Assisted Music Track

Jordan Ellis | Wallet: 0x7f3a...b291 | Domain: Music | Platform: SoundForge HAP

Design a Recipe

Jordan designs a reusable creative system: lo-fi hip hop, 70–85 BPM, minor key only, vinyl crackle texture required, melancholic tone, 4-layer maximum. Uploaded to IPFS → CID: bafybeigdyrzt5... Registered as: hap_recipe_v2_001 Established: A reusable, timestamped, owned creative system - before any track is made.

Register Voice

Jordan records vocal reference samples. Platform creates an encrypted voice model. Registered as: hap_voice_creator_jordy_001 Permissions: Music generation and remix permitted. Advertising, political, and adult content prohibited. Expires January 1, 2027. Revocable. Established: Vocal identity as a licensed asset generating royalty income independently of any track.

Provide Inputs

Lyric file v1 and compositional note: "Climax at 1:45; chorus vocal should feel distant, like a memory." Established: Explicit creative intent is documented and content-addressed.

AI Generation and Iteration

v1 - Rejected: "Tempo too fast, loses the late-night feel." v2 - Rejected: "Chorus needs to feel further away - more reverb, lower in the mix." v3 - Selected: "This one. The space in the chorus is right. Keep." Established: Rejection decisions document creative judgment - not just the final output.

HCS Calculation

Recipe (0.30) + Inputs (0.25) + Voice (0.20) + Iteration (0.15) + Curation (0.10) = 1.00 HCS = 0.90 → Tier: Primary Human Authorship

On-Chain Anchoring

1. Canonical JSON hashed: sha256:9f86d081... 2. Hash + wallet + timestamp written to Polygon smart contract 3. Transaction hash: 0xdef456... 4. Complete record uploaded to IPFS Established: Permanent, publicly verifiable proof of authorship - tamper-evident and independently verifiable.

Monetization

Three months later, a sync licensing platform finds the track: • Output license - rights to use the track in a film • Recipe license - a second producer licenses Jordan's recipe to make similar tracks • Voice license - a third party licenses Jordan's voice model for AI-generated harmonies All three traceable to a single HAP record.

Dispute

Another creator claims Jordan's recipe was derived from theirs (registered December 2024). Tier 1: Automated similarity check returns 71% - below 85% threshold. No hold. Tier 2: HAP Resolver reviews both records. Overlap attributed to shared genre conventions. Recommendation: Uphold Jordan's registration. Resolved in 18 days. Dispute record logged on-chain.

Formal Schema Specification

HAP v2.1 defines three canonical JSON schemas. Full machine-readable JSON Schema 2020-12 specifications are available at haphuman.xyz/schemas/v2.1/

15.1 Authorship Record Schema

The primary HAP record. Required fields: hap_version, record_id, domain_profile, hcs_components, contribution_score, hcs_tier, onchain_anchor, owner_identity.

record.json

{
  "$schema": "https://json-schema.org/draft/2020-12/schema",
  "$id": "https://haphuman.xyz/schemas/v2.1/authorship-record.json",
  "title": "HAP Authorship Record",
  "type": "object",
  "required": ["hap_version","record_id","domain_profile","hcs_components",
               "contribution_score","hcs_tier","onchain_anchor","owner_identity"],
  "properties": {
    "hap_version": { "type": "string", "const": "2.1" },
    "record_id": { "type": "string" },
    "domain_profile": { "enum": ["music","visual_art","written_content","custom"] },
    "contribution_score": { "type": "number", "minimum": 0, "maximum": 1 },
    "hcs_tier": {
      "enum": ["primary_human_authorship","collaborative_authorship",
               "ai_assisted_creation","ai_generated"]
    },
    "hcs_components": {
      "type": "object",
      "properties": {
        "recipe":    { "type": "number", "minimum": 0, "maximum": 1 },
        "inputs":    { "type": "number", "minimum": 0, "maximum": 1 },
        "voice":     { "type": "number", "minimum": 0, "maximum": 1 },
        "iteration": { "type": "number", "minimum": 0, "maximum": 1 },
        "curation":  { "type": "number", "minimum": 0, "maximum": 1 }
      }
    }
  }
}

15.2 Dispute Record Schema

Required fields: dispute_id, disputed_record_id, dispute_type, filed_by, filed_against, filed_at, status, tier.

record.json

{
  "$schema": "https://json-schema.org/draft/2020-12/schema",
  "$id": "https://haphuman.xyz/schemas/v2.1/dispute-record.json",
  "title": "HAP Dispute Record",
  "type": "object",
  "required": ["dispute_id","disputed_record_id","dispute_type",
               "filed_by","filed_against","filed_at","status","tier"],
  "properties": {
    "dispute_type": {
      "enum": ["ownership","derivation","voice_identity","contribution"]
    },
    "status": {
      "enum": ["pending","under_review","resolved","escalated","withdrawn"]
    },
    "tier": { "type": "integer", "minimum": 1, "maximum": 3 },
    "recommendation": {
      "enum": ["uphold","transfer","co_register","escalate","dismissed"]
    }
  }
}

15.3 Collaboration Record Schema

Extends the authorship record for multi-party works. The sum of all contributor contribution_share values must equal 1.0.

record.json

{
  "$schema": "https://json-schema.org/draft/2020-12/schema",
  "$id": "https://haphuman.xyz/schemas/v2.1/collaboration-record.json",
  "title": "HAP Collaboration Record",
  "allOf": [{ "$ref": "authorship-record.json" }],
  "properties": {
    "collaboration": { "type": "boolean", "const": true },
    "contributors": {
      "type": "array",
      "minItems": 2,
      "items": {
        "type": "object",
        "required": ["identity","role","hcs_components","contribution_share"],
        "properties": {
          "identity":           { "type": "string" },
          "role":               { "type": "string" },
          "contribution_share": { "type": "number", "minimum": 0, "maximum": 1 },
          "hcs_components":     { "type": "object" }
        }
      }
    },
    "attribution_agreement_cid": { "type": "string" }
  }
}

15b

Limitations

“HAP is a documentation and evidence framework. It does not create legal rights, guarantee copyright protection, or substitute for qualified legal counsel.”

What HAP Does Not Do

HAP cannot guarantee copyright protection for any work. No HCS score, tier, or on-chain record creates a copyright or establishes ownership as a matter of law. HAP does not replace legal counsel. Creators and platforms operating in regulated jurisdictions - particularly those handling voice data, offering revenue-sharing products, or asserting copyright - must engage qualified legal professionals. HAP does not create enforceable rights by itself. The framework documents creative decisions. The legal weight of that documentation depends on the jurisdiction, the specific facts, and applicable law at the time of any dispute. HAP records do not prove human agency. A HAP record documents what decisions were logged, not that a human (rather than an automated script) made those decisions. Platforms implementing HAP must design appropriate identity verification if human agency is material to their use case.

Version and Amendment Protocol

HAP records are interpreted under the schema version in effect at the time of creation. A record created under v2.1 is governed by v2.1 schema definitions and HCS weightings, regardless of subsequent version releases. When a new version is released, HAP will publish a migration guide specifying: which fields are backwards-compatible, which schema changes affect HCS calculation, and how dispute resolution handles cross-version record comparisons. Platforms must store the hap_version field and use it when interpreting archived records. Version-specific schema definitions are available at haphuman.xyz/schemas/v{version}/.

Future Directions

Standardization across platforms via open API specification and SDK
Integration with W3C Decentralized Identifiers (DIDs) for verified human identity binding
Legal alignment with evolving copyright frameworks across jurisdictions
Expansion of domain profiles beyond music, visual art, and written content
Algorithmic HCS validation tools for platform implementers
Cross-platform recipe and voice marketplaces

Conclusion

“Authorship is not defined by output alone. It is defined by the system that creates it.”

HAP enables a future where human creativity remains central, even in an AI-driven world. By making the process measurable, verifiable, and monetizable, we ensure that human contribution is never lost in the age of artificial intelligence.

Disclaimer

HAP is a documentation and attribution framework. It is not a legal service and does not provide legal advice. Nothing in this document constitutes legal, financial, or investment advice. HCS scores and tiers are descriptive indicators only. They do not constitute legal determinations of copyright eligibility, ownership, or authorship. Copyright eligibility is determined solely by applicable law. On-chain anchoring provides evidentiary support, not automatic legal priority or enforceable rights. Biometric and voice data collection is subject to complex and rapidly evolving federal, state, and international privacy law. The discussion in this document is illustrative only and does not constitute compliance guidance. Revenue-sharing and fractional ownership models may implicate securities law in certain jurisdictions. This document does not constitute securities advice. Creators, platforms, and investors should consult qualified legal counsel regarding copyright, privacy, securities, and regulatory compliance before implementing or relying on any aspect of this framework.

Join the Protocol

Human AuthorshipProtocol

Table of Contents

Abstract

Introduction

The Authorship Problem

2.1 Output vs Process

2.2 The Missing Layer

Defining Human Authorship

Core Components

4.1 Recipes (System Layer)

4.2 Voice (Identity Layer)

4.3 Inputs (Creative Layer)

4.4 Iteration (Process Layer)

4.5 Curation (Selection Layer)

Human Contribution Score (HCS)

5.1 What the HCS Measures

5.2 Weighting Rationale

Domain Profiles & HCS Tiers

Music (Default)

Visual Art

Written Content

Data Structure

Blockchain Integration

Stored On-Chain

Stored Off-Chain

The Anchoring Process

Ownership Model

8.1 Output Ownership

8.2 System Ownership

8.3 Identity Ownership

Monetization Model

Legal Framework

10.1 Copyright and AI-Generated Works

10.2 Jurisdiction Considerations

10.3 Voice Data and Biometric Privacy

10.4 HAP Records as Legal Evidence

10.5 Liability Boundaries

Off-Chain Data Integrity

11.1 Deletion and Privacy Rights

Dispute Resolution

Tier 1 - Automated Similarity Check (Platform Level)

Tier 2 - Structured Mediation (HAP Registry Level)

Tier 3 - External Arbitration

Anti-Gaming Protections

12.5 Dispute Record Schema

12.6 Collaboration and Split Attribution

Relationship to C2PA

Worked Example - AI-Assisted Music Track

Design a Recipe

Register Voice

Provide Inputs

AI Generation and Iteration

HCS Calculation

On-Chain Anchoring

Monetization

Dispute

Formal Schema Specification

15.1 Authorship Record Schema

15.2 Dispute Record Schema

15.3 Collaboration Record Schema

Limitations

What HAP Does Not Do

Version and Amendment Protocol

Future Directions

Conclusion

Disclaimer

Human Authorship
Protocol