Copyright Implications Of AI-Generated Sculpture Curation.
1. Authorship of AI-Generated Music from Genetic Data
Core Issue:
If AI generates music using a person’s genetic sequence as input, who owns the copyright?
Is the AI author?
Is the human who provided genetic data the author?
Is the AI programmer or curator the author?
🔹 1.1 Thaler v. Perlmutter (United States)
Facts:
Stephen Thaler registered AI-generated artwork for copyright with the AI itself as the author.
Holding:
Only human authorship is recognized under U.S. copyright law.
AI alone cannot be an author.
Implication for AI-Generated Music:
AI-generated music from genetic data cannot be copyrighted if entirely autonomous.
If a human composes or guides the AI, they may claim authorship.
🔹 1.2 Naruto v. Slater
Facts:
A monkey took a selfie with a photographer’s camera.
Court ruled non-human creators cannot own copyright.
Implication:
Like animals, AI lacks legal personhood.
Music generated from genetic sequences solely by AI is likely public domain in the U.S.
2. Originality & Creativity in AI-Generated Music
🔹 2.1 Feist Publications v. Rural Telephone Service
Principle:
Copyright requires independent creation + minimal creativity.
Application:
AI output alone may not satisfy originality.
Human curation of AI music (selecting motifs, editing sequences, mapping genetic patterns creatively) may meet minimal creativity.
For interactive exhibitions or albums derived from genetic sequences:
The human arranger’s contribution is crucial to obtain copyright protection.
🔹 2.2 Computer Associates International v. Altai
Facts:
Dispute over computer software copying.
Principle:
Abstraction-Filtration-Comparison (AFC) test isolates protectable expression from ideas or processes.
Application:
AI music generated from genetic code is similar:
Genetic data = idea/methodology (not protectable)
Resulting musical pattern = expression (potentially protectable if human-influenced).
3. Derivative Works and Training Data
AI music often uses pre-existing music for style transfer or training.
🔹 3.1 Andersen v. Stability AI
Facts:
Artists sued AI companies for training on copyrighted works.
Legal Questions:
Does AI training constitute copyright infringement?
Are outputs derivative works of copyrighted material?
Relevance:
Genetic-music AI models trained on existing compositions may infringe if outputs are substantially similar.
Exhibitions displaying these works could be liable.
🔹 3.2 Authors Guild v. Google
Facts:
Google scanned copyrighted books to create a searchable database.
Holding:
Court allowed it under fair use, emphasizing transformative purpose.
Application:
If AI training is transformative, e.g., turning DNA sequences into original musical compositions, it may be defended as fair use.
But commercial sale or exhibition could complicate defense.
4. AI-Curated Sculpture Exhibitions
AI curation involves selecting, arranging, and sometimes modifying sculptures for display.
🔹 4.1 Rogers v. Koons
Facts:
Jeff Koons copied a photograph into a sculpture.
Court found infringement: substantial similarity = violation.
Application:
AI curation that replicates identifiable sculptures may infringe.
AI cannot claim originality if it outputs a derivative copy.
🔹 4.2 MAI Systems Corp. v. Peak Computer
Principle:
Fixation requirement: software loaded into RAM is “fixed” for copyright purposes.
Application to Sculpture Curation:
Digital representations of sculptures (3D scans, AI-generated modifications) are fixed works, copyrightable if human contribution exists.
🔹 4.3 Baker v. Selden
Principle:
Copyright protects expression, not ideas, methods, or systems.
Application:
AI-curated exhibitions’ algorithmic selection method is not copyrightable.
Only final arrangement or modifications made by humans can be protected.
5. Moral Rights and Attribution
Jurisdictions like EU countries recognize moral rights, which protect the author’s reputation and prevent distortion.
AI curation altering an artist’s work (e.g., changing style or placement in exhibition) could violate moral rights even if AI-generated.
6. Comparative Jurisdiction Notes
| Issue | U.S. | UK | EU |
|---|---|---|---|
| AI authorship | Not recognized (Thaler) | Recognized if human made arrangements | Similar to UK, with stronger moral rights |
| Derivative work from training | Depends on fair use (Google) | More restrictive | Often strict; moral rights included |
| Genetic-data music | No copyright unless human involvement | Could qualify if human input | May qualify, but moral rights can apply |
| Curation | Fixation + human contribution | Protectable if human-guided | Stronger protection; moral rights + attribution |
7. Key Takeaways
AI alone cannot hold copyright (Thaler, Naruto).
Human creative input is essential for both music and sculpture curation (Feist, Computer Associates).
Derivative work risks exist if AI uses copyrighted training data (Andersen v. Stability AI, Rogers v. Koons).
Transformative AI outputs may qualify for fair use (Authors Guild v. Google).
Fixation and tangible form still matters for copyright (MAI Systems).
Moral rights in some jurisdictions may impose additional restrictions (EU law).
✅ Practical Implications
Curators should document human contributions (selection, editing, mapping).
AI training datasets should avoid infringing copyrighted works.
Legal strategies should consider jurisdictional differences.
Genetic-data-based music may be public domain in the U.S. unless meaningful human creative input exists.
RELATED Blog
comments