THE CIP STANDARD
The framework for subsisting AI⇆IP rights for creative intellectual property. Developed by Creative Intellectual Property Charity with input from creators, agencies, lawyers, underwriters, and platform operators.
| Document | The CIP Standard v2.1 |
| Status | Adopted by Trustee Board |
| Published | November 2025 |
| Last revised | April 2026 |
The CIP Standard
A framework for subsisting AI⇆IP rights for creative intellectual property — provenance, consent, governance, and professional recognition across the AI content pipeline.
The CIP Standard sets out a clear framework for creative content provenance, rights awareness, consent, governance, and professional recognition. It exists because creative works carry subsisting IP rights that persist across every form of AI use — from training data ingestion to synthetic content generation to automated distribution. The standard provides a shared structure for creators, organisations, advisers, and platforms working to understand and professionally manage those rights.
Preliminary: the renaming problem
The core problem is definitional. When a creative work is renamed “training data”, the rights that subsist in it do not disappear. Copyright does not lapse. Neighbouring rights are not waived. Identity rights are not surrendered. Yet the operational frameworks governing AI training pipelines have treated those rights as if they had.
CIP exists to correct that: to bring structure, professional standards, and recognised evidence to the question of what rights subsist, in what content, and on what terms AI use is or is not permitted.
A common standard helps create consistency. It gives creators, organisations, advisers, and platforms a shared basis for understanding what is expected, what evidence matters, and what recognised good practice looks like.
Subsisting rights and registered rights
This distinction is fundamental to understanding why the CIP framework exists and what problem it solves.
Subsisting rights — automatic legal recognition
A subsisting right exists the moment a qualifying work is created or a qualifying act occurs. No application. No registration. No fee. No government office involved. The right comes into existence automatically under law.
- Copyright
- Subsists from creation in songs, photographs, writing, and other original works. No application required.
- Moral rights
- Subsist alongside copyright, protecting the author’s right to attribution and the integrity of their work.
- Neighbouring rights
- Subsist in a performance or recording from the moment it is fixed. Protect performers and producers.
- NILP rights
- Name, image, likeness, and persona rights subsist as commercial identity markers. The fastest-moving area of AI law.
- Database rights
- Subsist in structured datasets where sufficient investment was made in their creation.
- Biometric rights
- Subsist in faceprints, voiceprints, and other biological identifiers used by AI systems.
Registered rights — state-issued exclusivity
A registered right is granted by a state authority following a formal application and examination process. Registration requires active steps by the rights holder and, typically, payment of fees. Examples include patents, trade marks, and registered designs.
Why this distinction matters for AI
The AI content pipeline almost exclusively engages subsisting rights, not registered rights. When an AI company scrapes the web, ingests a music catalogue, or trains on photographs, the rights it encounters are automatic and universal — they do not depend on whether any creator filed a patent or registered a trade mark.
This is why renaming creative content as “training data” does not help AI companies legally. The copyright subsists. The neighbouring rights subsist. The NILP rights subsist. None of them required registration. None of them lapse because the work was ingested.
The six functional areas
The CIP Standard brings together core areas that increasingly shape the handling of creative content. It is organised around distinct functions in the creative rights chain.
- Provenance
- Where content comes from and how it is evidenced.AI⇆IP relevance: Establishing chain of origin before AI ingestion.
- Rights awareness
- The rights that may exist in a work, asset, performance, likeness, or dataset.AI⇆IP relevance: Identifying subsisting rights before and after AI use.
- Consent
- How permission is granted, limited, reviewed, or withdrawn.AI⇆IP relevance: Consent to AI training; temporal and scope-limited licences.
- AI pipeline exposure
- How content moves through AI training, generation, and distribution stages.AI⇆IP relevance: The specific pipeline stages where subsisting rights are engaged.
- Governance
- Responsibilities, controls, and oversight across participants.AI⇆IP relevance: Accountability for rights-aware ingestion and output.
- Recognition
- Credentials, marks, and formal recognition against defined standards.AI⇆IP relevance: Verified trust signals for rights-compliant AI operations.
Machine-readable declaration: cip.md
The cip.md file is the machine-readable expression of the CIP Standard. It sits alongside creative content — in the same directory, attached to the same asset, deployed at the same endpoint — and declares the rights, consent terms, and licensing conditions that apply to the work.
The format is designed to be readable by both humans and AI systems. It uses a structured key-value syntax that can be parsed programmatically while remaining intelligible to anyone who opens the file. The parallel with robots.txt is deliberate: a lightweight, file-based declaration that AI systems are expected to read and respect.
CIP exists precisely because subsisting rights are invisible to AI systems that have not been designed to look for them. A CDR makes subsisting rights visible. A cip.md declaration makes them machine-readable. The Rights Registry makes them verifiable. The certification framework ensures that the professionals handling these rights understand, evidence, and respect them.
How the standard is organised
The CIP Standard is organised through role-based certification paths. Each path maps to a distinct function in the rights chain and has its own requirements, assessments, standards, and outputs.
- Creator Certification
- Individuals who originate rights-bearing content.
- Agency Certification
- Organisations managing rights across client portfolios.
- Legal Practitioner Certification
- Legal professionals advising on rights and regulation.
- Underwriter Certification
- Underwriters and risk professionals.
- Platform Certification
- Platforms, distributors, broadcasters, and brands.
Each pathway is separate. Recognition in one path does not replace the requirements of another.
Governance and oversight
CIP is governed by Creative Intellectual Property Charity through a standards structure with representation from the professional communities it serves. Oversight includes standards governance, independent audit support, and ongoing review as regulation and professional practice develop.
Versioning and development
The CIP Standard is currently at version 1.0. As legal frameworks, professional practice, and sector expectations develop — particularly around AI training obligations, synthetic content disclosure, and identity rights — the standard will be revised, expanded, and updated. Versioning ensures changes are structured, visible, and reviewable.
This document is reviewed on a 6-monthly cycle by the Standards Committee. The next scheduled review is October 2026.
Annotations in this document follow the CIP convention: marginal notes provide context; they do not form part of the normative standard text.
Consultation responses and feedback on the standard should be directed to charity@creativeip.org.
- CIP Level 1 Foundation Course — 7 modules covering the fundamentals
- Certification Tracks — 5 professional certification paths
- CIP Glossary — 94 defined terms across three editorial layers
- AI IP Risk Self-Assessment — 8 questions, 5 minutes, actionable results
THE FRAMEWORK EXISTS.
The standard is published. The courses are open. The tools are available. Start with the self-assessment to understand your rights exposure, or go straight to the Level 1 course.