Copyright Implications For AI-Assisted Ethnomusicology And Cultural Heritage Preservation
I. Core Copyright Issues in AI-Assisted Ethnomusicology
1. Authorship & Ownership
Who owns AI-generated transcriptions of traditional songs?
Is AI a “creator” under copyright law?
Does the researcher, the programmer, or the community hold rights?
Most jurisdictions require human authorship.
2. Public Domain vs. Traditional Cultural Expressions (TCEs)
Many traditional songs:
Are centuries old
Have no identifiable author
Fall into the public domain
However:
Indigenous communities may claim cultural ownership
Moral and collective rights may still exist
Copyright law often fails to fully protect communal heritage.
3. Fixation & Originality
When AI:
Cleans up old recordings
Enhances field tapes
Reconstructs incomplete melodies
Generates “new” works in traditional style
The legal question becomes:
Is this a new original work, or just a mechanical reproduction?
4. Derivative Works
AI-trained on:
Archival recordings
Sacred songs
Community performances
If it generates similar outputs:
Is that infringement?
Or fair use?
Or transformative?
5. Moral Rights & Cultural Integrity
In some jurisdictions (e.g., France, India):
Moral rights protect integrity of works.
Distortion of sacred songs may violate rights even if copyright expired.
II. Detailed Case Law Analysis
1. Feist Publications, Inc. v. Rural Telephone Service Co.
Facts:
Rural Telephone compiled a phone directory. Feist copied listings. Rural claimed copyright.
Legal Issue:
Does mere compilation of factual information qualify for copyright?
Holding:
The U.S. Supreme Court ruled:
Copyright requires originality
Mere “sweat of the brow” is insufficient
Minimal creativity is required
Relevance to AI Ethnomusicology:
If an AI:
Transcribes folk songs
Organizes them into a database
Digitizes field recordings
That database is protected only if:
There is creative selection or arrangement
Simple transcription of public domain folk songs → not protected.
But:
Curated thematic arrangement of indigenous music traditions → potentially protected.
This case establishes the minimum creativity threshold.
2. Naruto v. Slater
Facts:
A macaque monkey took a selfie using a photographer’s camera. PETA argued the monkey owned the copyright.
Legal Issue:
Can a non-human be an author?
Holding:
Only humans can hold copyright under U.S. law.
Relevance to AI:
If:
AI generates reconstructed tribal music
AI creates new compositions in indigenous style
AI cannot be the author.
Therefore:
Ownership must vest in:
The human operator
Or possibly no one (if no human creativity involved)
This case strongly influences global debates on AI authorship.
3. Bridgeman Art Library v. Corel Corp.
Facts:
Corel reproduced exact photographic copies of public domain paintings. Bridgeman claimed copyright in the photos.
Holding:
Exact photographic reproductions of public domain works lack originality.
Relevance:
If AI:
Restores an old field recording of a 19th-century tribal chant
Faithfully reproduces it without creative alteration
Then:
The restoration may not qualify as a new copyrighted work.
This case limits claims over digitized heritage material.
4. Eastern Book Company v. D.B. Modak
Facts:
The issue concerned copyright in edited versions of Supreme Court judgments.
Holding:
India adopted a “modicum of creativity” standard (not merely labor).
Relevance:
If Indian researchers:
Use AI to annotate and arrange traditional ragas
Create structured notational archives
Protection requires creative editorial contribution, not mere digitization.
This case is critical for AI-driven ethnomusicology in India.
5. Infopaq International A/S v. Danske Dagblades Forening
Facts:
Infopaq copied 11-word extracts from newspapers for data analysis.
Holding:
Even small parts of a work can be protected if they reflect author’s intellectual creation.
Relevance:
If AI:
Extracts short melodic fragments
Uses rhythm patterns from protected recordings
Even small segments may infringe if recognizable.
Important for:
AI training datasets containing copyrighted performances.
6. Andy Warhol Foundation v. Goldsmith
Facts:
Warhol used photographer Goldsmith’s image of Prince to create stylized artwork.
Holding:
Commercial licensing that competes in the same market may not qualify as fair use, even if stylistically different.
Relevance:
If AI:
Generates music “in the style of” a contemporary indigenous musician
Sells it commercially
Even transformative style changes may not guarantee fair use.
This narrows AI reliance on “transformative use.”
7. Authors Guild v. Google, Inc.
Facts:
Google scanned millions of books to create a searchable database.
Holding:
Digitization for search and indexing was fair use because it was highly transformative.
Relevance:
AI training for:
Pattern recognition
Musical transcription
Academic research
May qualify as fair use if:
Non-commercial
Transformative
Does not substitute the original market
This case supports archival digitization projects.
8. Thaler v. Comptroller-General of Patents, Designs and Trade Marks
Facts:
Stephen Thaler claimed AI (DABUS) invented a patentable invention.
Holding:
Only natural persons can be inventors under UK law.
Relevance:
Reinforces that:
AI cannot be legal author or inventor.
Thus, AI-generated folk reconstructions without human creative input may fall into legal uncertainty.
III. Cultural Heritage and Indigenous Rights Dimension
Beyond copyright:
International frameworks like:
UNESCO cultural heritage conventions
WIPO discussions on Traditional Cultural Expressions (TCEs)
Recognize that:
Even public domain works may require community consent.
AI complicates this by:
Scaling reproduction
Enabling stylistic imitation
Removing community control
Copyright law often protects individual authors, while ethnomusicology deals with collective heritage.
IV. Key Legal Tensions
| Issue | Legal Position | Ethnomusicology Impact |
|---|---|---|
| AI Authorship | Human required | AI outputs uncertain |
| Public Domain Folk Music | Free to use | Ethical concerns remain |
| Restoration of Archives | Needs creativity | Mechanical restoration unprotected |
| AI Training on Recordings | May be fair use | Risk of infringement |
| Sacred Songs | Weak copyright protection | Cultural rights gap |
V. Emerging Legal Questions
Should communities have collective copyright?
Should AI-generated cultural outputs be restricted?
Can moral rights extend to traditional groups?
Should AI training require cultural consent?
These remain unresolved globally.
Conclusion
AI-assisted ethnomusicology sits at the intersection of:
Copyright law
Indigenous cultural rights
Moral rights doctrine
Public domain principles
AI authorship debates
The major cases above collectively establish that:
Originality requires human creativity
Exact reproductions lack protection
Small fragments can infringe
AI cannot be an author
Transformative use has limits
Digitization for research may be fair use
However, copyright law remains structurally designed for individual authorship, not communal cultural heritage. This creates a regulatory gap when AI engages with traditional music traditions.
RELATED Blog
comments