Copyright Implications For Neural AI Collaborative Visual Art Projects.
I. Legal Context
Neural AI collaborative visual art projects involve:
Neural networks or AI systems generating images, paintings, or multimedia art
Human-AI collaboration, where humans provide prompts, select outputs, or edit AI-generated visuals
Key Philippine and international legal principles:
Republic Act No. 8293 protects literary and artistic works that are original intellectual creations.
Human authorship is required. AI alone cannot be an author.
Moral rights, including attribution and integrity, belong to humans.
Copyright arises upon creation, not registration.
II. Key Legal Issues
Authorship and Ownership
Who qualifies as the author in a human-AI collaborative project?
Can AI-generated visual art without human intervention be copyrighted?
Originality and Creativity
Does the output meet the threshold of original intellectual creation?
How much human contribution is necessary?
Derivative Works
If AI is trained on copyrighted art, does creating new images infringe existing rights?
Moral Rights
Does collaboration affect the right to attribution and integrity?
III. Relevant Case Laws
1. Feist Publications, Inc. v. Rural Telephone Service Co. (US Supreme Court, 1991)
Facts:
Feist copied Rural Telephone’s phone listings.
Holding:
Facts are not copyrightable; only original selection or arrangement qualifies.
Relevance:
AI-generated art based on learned patterns from preexisting works may not be original.
Human selection, arrangement, or editing of AI outputs may satisfy originality requirements.
2. Burrow-Giles Lithographic Co. v. Sarony (US Supreme Court, 1884)
Facts:
A photograph of Oscar Wilde was copied without permission.
Holding:
Copyright protects works with creative human control, such as composition, lighting, or pose.
Relevance:
Human contributors who craft prompts, select AI outputs, and refine visuals may qualify as authors.
Pure AI creations without meaningful human input are likely not protected.
3. Naruto v. Slater (US Court of Appeals, Ninth Circuit, 2018)
Facts:
A macaque monkey took a selfie; a group tried to claim copyright on its behalf.
Holding:
Non-human entities cannot hold copyright.
Relevance:
AI systems, like animals, cannot be considered authors.
Purely AI-generated visual art without human authorship cannot be copyrighted.
4. Thaler v. Vidal (US, 2023, cert denied)
Facts:
Stephen Thaler tried to register AI-generated artwork, listing the AI as author.
Holding:
Copyright requires human authorship.
Registration of purely AI-generated works was denied.
Implications:
Neural AI collaborative projects require human intellectual contribution to be protected.
5. Aalmuhammed v. Lee (US Court of Appeals, Ninth Circuit, 2000)
Facts:
A consultant claimed co-authorship of the film Malcolm X.
Holding:
Authorship requires control over the creative process, supervision, and intellectual contribution.
Relevance:
In human-AI art, the human who guides the AI, selects outputs, and edits images may be recognized as the author.
6. Infopaq International A/S v. Danske Dagblades Forening (CJEU, 2009)
Facts:
Infopaq copied newspaper content for indexing.
Holding:
A work is protected if it reflects the author’s own intellectual creation, including selection, coordination, and arrangement.
Relevance:
Human creative input in prompt design, image selection, and composition may render AI-assisted visual art copyrightable.
Fully autonomous AI output may not meet this standard.
7. Authors Guild v. Google, Inc. (US Court of Appeals, Second Circuit, 2015)
Facts:
Google scanned books to create a searchable database.
Holding:
Use was transformative fair use, even though it reproduced copyrighted text.
Relevance:
Using AI-trained models on copyrighted artworks may be acceptable for transformative purposes, such as research or non-commercial art generation.
Commercial reproduction of AI-generated art closely mimicking copyrighted works may still be infringing.
8. Bridgeman Art Library v. Corel Corp. (US District Court, 1999)
Facts:
Corel reproduced exact digital photographs of public domain artworks.
Holding:
Exact reproductions without creative input are not copyrightable.
Relevance:
AI reproductions of existing artworks (even modified) may not qualify unless the human author applies substantial creative input.
IV. Philippine-Specific Considerations
Human Authorship Requirement
Neural AI cannot be the author.
Human authorship is required to claim copyright.
Derivative Works
AI trained on copyrighted works may generate derivative content.
Unauthorized reproduction could infringe moral and economic rights.
Moral Rights
Right to attribution and integrity may extend to the human collaborator who guided or refined the AI outputs.
Collaborative Projects
Multiple humans working with AI can jointly claim authorship if each makes an intellectual contribution.
V. Ownership Scenarios
| Scenario | Copyright Implication |
|---|---|
| Fully autonomous AI-generated art | Not copyrightable |
| Human provides prompts, selects, and edits AI output | Human owns copyright |
| Multiple human collaborators guiding AI | Joint authorship possible |
| AI reproduces existing copyrighted works | May infringe unless transformative or fair use |
VI. Infringement Risks
Copying or closely imitating copyrighted art in AI training or outputs.
Claiming AI as author (not legally valid).
Misattribution that harms human moral rights.
VII. Key Takeaways
Human authorship is essential: AI alone cannot be an author (Naruto v. Slater, Thaler v. Vidal).
Originality requires human contribution: Prompting, selection, editing, and composition are key (Burrow-Giles, Aalmuhammed).
Derivative content risk: AI trained on copyrighted material may produce infringing works.
Moral rights apply: Human collaborators maintain rights of attribution and integrity.
Fair use may protect research or non-commercial use (Authors Guild v. Google).
VIII. Conclusion
Neural AI collaborative visual art projects exist at the intersection of technology, creativity, and law. The prevailing legal trend emphasizes:
AI cannot hold copyright
Human creativity is the key to protection
Fully autonomous AI works are not protected
Moral and derivative rights are enforceable against misuse
Case law from Feist, Burrow-Giles, Naruto, Thaler, Aalmuhammed, Infopaq, Authors Guild, and Bridgeman collectively support these principles.
RELATED Blog
comments