Source material declaration — third-party content in the operator's asset.

Many operator assets contain third-party material — quotations, samples, clips, archival footage, stock photos, library music, datasets, code components — present in the asset but owned by parties other than the operator. The framework's existing rights infrastructure addresses how the operator's own content is rights-positioned; the Source Material Declaration architecture extends to declaring third-party material the operator has incorporated. v3.34 ships the architectural anchor and the generalised CIP-Source-Material field as the starter increment; nine category workstreams are named and queued for subsequent releases with dedicated design-decision turns.

Why source material warrants cross-cutting architectural treatment

Source material is sector-crosscutting by nature. Documentary filmmakers using archival footage; journalists using quoted speech and excerpted reports; academic writers using quotations and structured datasets; music producers using samples and library music; game developers using third-party assets and open-source code; advertising agencies using stock photography and library music; podcasters using clips and quoted excerpts; visual artists using found imagery and reference material — every creative sector engages source-material declaration. Treating it as a sector-vertical concern would force premature specialisation; treating it cross-cuttingly preserves architectural coherence.

The framework's existing infrastructure addresses adjacent but distinct concerns. CIP-Mixed-Rights-Block (v3.4) handles per-component rights for multi-component operator-controlled assets — a film with multiple rights-bearing layers the operator administers. CIP-Third-Party-Marks (v3.33) addresses third-party trademarks specifically, distinct from copyrighted material. CIP-Input-Licence declares how the operator came to hold the asset as a whole. None of these handles the case where the operator's content includes specific identifiable third-party copyrighted material that the operator has incorporated under licence, fair use, public-domain status, or other clearance basis.

The operational significance of declaring source material: AI tooling vendors and downstream platforms encountering an asset with declared source material know that AI uses of that asset must respect the source material's rights independently of the operator's own rights position. An AI training pipeline ingesting content with declared unauthorised source material should treat that as a flag requiring separate clearance analysis. An AI generation system producing derivatives should similarly flag declared source material for downstream review. A documentary filmmaker who has cleared archival footage for the original release but not for AI training derivative use can declare this position machine-readably rather than relying on the absence of an affirmative permission.

The nine source-material categories

The architecture identifies nine distinct categories of source material, each with characteristic permission patterns, industry infrastructure, and jurisdictional variation. v3.34 names and queues each as a workstream for subsequent releases.

Category 1 — Quotations and excerpts

Text from books, articles, speeches, interviews, court judgments, broadcast transcripts. Used in journalism, academic writing, criticism, biography, documentary commentary. Permission bases vary widely: licensed (rights-holder permission), fair-use / fair-dealing (criticism, review, news reporting, research), public-domain (works whose copyright has expired), or attribution-only (where the source permits unconditional use). Operationally significant for content claiming critical-commentary fair-use posture against AI training (the academic-writer-on-copyright case).

Category 2 — Audio samples

Recorded audio fragments incorporated into the operator's audio production. Music production sampling, vocal samples, sound effects libraries, found-sound. Permission bases concentrate on licensed (sample clearance), creative-commons (CC0 or CC-BY samples from libraries like Freesound), or public-domain (early recordings now out of copyright). The sample-clearance vertical has substantial commercial infrastructure (sample clearance services, sample-clearance houses) the framework cross-references rather than duplicating.

Category 3 — Film and TV clips

Audiovisual clips from films, TV programmes, broadcasts incorporated into documentaries, news segments, criticism content, educational material, parody, or compilation works. Permission bases include licensed (clip licensing, often festival-only or limited-territory), fair-use (US documentary fair-use jurisprudence including Joy of Music v. Hugo Music; UK fair-dealing for criticism / review / news reporting), or cleared-other (where unusual clearance arrangements apply — bespoke documentary releases, archival licence terms). Documentary filmmaking has a substantial fair-use practice that the framework's fair-use permission status accommodates declaratively.

Category 4 — Archival material

Historical footage, period photographs, news footage, broadcast archives, government records, military / institutional archives. Permission bases vary substantially by archive: licensed (commercial archival licence — Getty Archive, AP Archive, Reuters, BFI National Archive, BBC Archive), public-domain (works in the public domain by age or status — pre-1929 US works, UK government works after the relevant Crown Copyright period), creative-commons (some institutional archives release material under CC), or cleared-other (museum-loan terms, foundation-licensed material). Operationally significant for historical documentary, period drama productions, museum-curated content, educational material.

Category 5 — Stock material

Stock photography, stock footage, library music, stock illustrations, royalty-free assets. Common in advertising, web content, podcasts, business video, social media. Permission bases concentrate on licensed (commercial stock licence — Getty Images, Shutterstock, Adobe Stock, Epidemic Sound, Artlist, Audioblocks), creative-commons (Wikimedia, Pexels, Unsplash, Pixabay materials under various CC variants), or editorial-use-only (a common stock subset where commercial use is excluded by licence terms). The “editorial-use-only” pattern is operationally significant because it engages a distinct AI-use question — AI training on editorial-use-only material requires separate analysis.

Category 6 — Performances

Third-party performances incorporated into the operator's asset — cover songs, sampled performances, licensed live recordings, archival performance footage. Permission bases include licensed (sync licence, master use licence, performance rights licence), creative-commons (where the performer has CC-licensed their performance), public-domain (where the performance is in the public domain by age — historic performances now out of copyright), or cleared-other. Engages neighbouring rights for performers (UK CDPA Part II, US Sound Recording Performance Right, EU Rental and Lending Rights Directive). Cross-references the v3.32 Film & TV Vertical Architecture's Workstream 4 (talent agreement integration) where the performance source material engages the operator's own talent agreements.

Category 7 — Quoted dialogue and characters

Fictional quotation — characters referenced or evoked, dialogue quoted, fictional-world incorporation. Common in fan fiction, parody, critical commentary, transformative works. Permission bases concentrate on fair-use / fair-dealing (parody, criticism, transformative use under Campbell v. Acuff-Rose), licensed (where the rights-holder has authorised character use — sanctioned fan fiction, official tie-ins), creative-commons (rare but exists for some fictional universes), or unauthorised. The character-rights area is jurisdictionally complex (US character copyright doctrine vs UK character protection through copyright + passing-off + trademark).

Category 8 — Datasets and structured material

Licensed datasets, structured data resources, statistical compilations used as part of the operator's asset. Common in data journalism, academic publication, business intelligence content, AI training datasets used as input to operator's AI generation work. Permission bases include licensed (commercial data licence — Bloomberg, FactSet, academic-database licences), creative-commons (open data — most government open-data programmes, Wikimedia structured data), public-domain (US government data, certain UK Crown Copyright after expiry), or cleared-other. Engages Database Rights (UK CDPA Part III, EU Database Directive 96/9/EC) which subsist independently of copyright.

Category 9 — Software and code

Third-party software libraries, open-source incorporation, licensed code components present in the operator's asset where the asset includes code (software products, interactive content, websites, games). Permission bases concentrate on creative-commons or specific open-source licences (MIT, Apache 2.0, GPL family, BSD family — recognised through the standard SPDX identifier vocabulary), licensed (commercial library licence), or cleared-other. The Software Bill of Materials (SBOM) practice from the cybersecurity domain provides parallel infrastructure the framework cross-references rather than duplicating.

The CIP-Source-Material starter field (v3.34)

v3.34 ships a single generalised field that operators can use to declare source material across any of the nine categories. The field follows the v3.33 CIP-Third-Party-Marks pattern — multi-entry comma-separated, structured sub-components, closed permission-status vocabulary — extended with three additional permission-status values that source material engages but third-party marks typically don't: public-domain, creative-commons, and cleared-other.

CIP-Source-Material [mode: dual]— declares third-party content (other than trademarks, which are handled by CIP-Third-Party-Marks) present in the operator's asset. Format: {source-description}:{rights-holder}:{permission-status}, comma-separated multi-entry support. Dual-mode classification: operator publishes the declaration (steering); the underlying rights position is recorded by the third-party rights-holder or its licensing infrastructure (recording from their perspective).

Source description

Free-text identifier for the source material. Should be specific enough to identify the material if challenged. Examples: Reuters-archive-clip-20230315, Getty-photo-1742893, Beatles-Strawberry-Fields-sample-12s, BFI-archive-WWII-footage-1944-Calais, NYT-1985-article-excerpt, Shakespeare-Hamlet-Act3-Scene1. Free-text rather than closed vocabulary because source-description varies materially by category.

Rights holder

The legal entity holding the rights to the source material. Examples: Reuters Limited, Getty Images Inc., Sony Music Entertainment, British Film Institute, The New York Times Company, Public Domain (where applicable). For complex rights structures (where master and publishing rights sit with different parties), the field may carry multiple holders separated by semicolons within the rights-holder sub-component (e.g., Sony Music; Sony Music Publishing).

Permission status

Closed-vocabulary value declaring the operator's basis for inclusion. Seven values:

• licensed — operator has an explicit licence from the rights-holder. Implicit acknowledgement: operator can substantiate the licence terms (territorial scope, term, modifications permitted, AI use carve-outs) if challenged.
• fair-use— operator relies on fair use (US 17 U.S.C. §107) or fair dealing (UK CDPA §§29–30) or equivalent jurisdictional exception. Per-jurisdiction analysis essential.
• incidental— inclusion is incidental and the operator's position is that no licence is required (UK CDPA §31 incidental inclusion exception and analogues). Operationally distinct from fair-use.
• public-domain— work is in the public domain in the operator's jurisdiction. Operator must substantiate (UK life+70; US per 17 U.S.C. §§302–305; Crown Copyright variations).
• creative-commons— work is licensed under a Creative Commons licence variant. Operator's notes carry the specific variant (CC0, CC-BY, CC-BY-SA, CC-BY-NC, CC-BY-ND, CC-BY-NC-SA, CC-BY-NC-ND) — variant materially affects permitted uses including AI training.
• cleared-other — cleared through arrangements not fitting other categories (museum-loan terms, foundation-licensed, festival-only documentary licences, research-use-only datasets).
• unauthorised — declared for evidentiary purposes. Distinct from concealment — declaring unauthorised is itself a transparency act.

Worked example

CIP-Source-Material: BFI-archive-WWII-footage:British Film Institute:licensed,
                     Reuters-clip-20230315:Reuters Limited:licensed,
                     Churchill-1940-speech-excerpt:Public Domain:public-domain,
                     Shakespeare-Hamlet-quotation:Public Domain:public-domain,
                     interview-Smith-2022:Smith Estate:licensed,
                     Beatles-Strawberry-Fields-2s-clip:Sony Music; Sony Music Publishing:fair-use

Six different source-material entries spanning archival film, news clips, public-domain speech, public-domain literary text, licensed interview, and a fair-use audio sample. All declared in a single field value with multi-entry comma-separated support.

Disposition summary — workstreams queued

v3.34 ships the generalised CIP-Source-Material field as the starter increment. The nine category workstreams (quotations, samples, clips, archival, stock, performances, dialogue and characters, datasets, software and code) are named and queued for subsequent releases. Each warrants its own design-decision turn before authoring category-specialised fields. The framework's discipline established at v3.32 (Film & TV Vertical Skeleton-First) of asking the user for substantive design choices before authoring is preserved.

Specialised field decisions deferred: whether the framework needs CIP-Source-Audio-Sample with sample-specific structured sub-components (timecode, sample-rate, master-vs-publishing-split) versus continuing to use the generalised field; whether CIP-Source-Stock warrants a dedicated field referencing standard stock-licence taxonomies (standard-licence vs extended-licence vs editorial-only); whether CIP-Source-Code warrants integration with the SPDX licence identifier vocabulary; and so on. Each specialisation creates field-set bloat and warrants justification before adding.

Cross-references

This architecture cross-references existing framework material. The Mixed-Rights Architecture (v3.4) handles per-component rights for multi-component operator-controlled assets — distinct from but adjacent to source material declaration. The cip.md Specification documents the field set including the v3.33 CIP-Third-Party-Marks for trademarks specifically and the v3.34 CIP-Source-Material for everything else.

External references (verification caveat applies): Creative Commons licence variants (CC BY 4.0, CC BY-SA 4.0, CC0 1.0, etc.) maintained at creativecommons.org; SPDX licence identifier vocabulary maintained at spdx.org/licenses; UK CDPA 1988 Parts II (performances) and III (database right); EU Database Directive 96/9/EC; US 17 U.S.C. §§107 (fair use), 302–305 (copyright term); Crown Copyright provisions per UK government guidance.

This document: CIP Source Material Declaration Architecture v3.34 · April 2026. Authority: Creative Intellectual Property Charity Standards Committee. Status: Skeleton with the generalised CIP-Source-Material field as starter increment; nine category workstreams named and queued. Verification caveat: source-material licensing references and per-jurisdiction analyses are draft from training memory; qualified entertainment, media, or IP counsel should review before treatment as authoritative. Licence:© Creative Intellectual Property Charity. All Rights Reserved.