Ipr In AI-Generated Digital Art
1. Introduction to IPR in AI-Generated Art
AI-generated art refers to works created either entirely or partially by artificial intelligence, without direct human authorship. This raises unique challenges for IPR because:
Copyright laws traditionally protect human authors.
Ownership and authorship become unclear when AI autonomously generates works.
Questions arise about whether AI itself can hold copyright, or if it should be assigned to the programmer, user, or neither.
Key legal issues include:
Authorship: Who is the "author" of AI-generated art?
Ownership: Who owns copyright—user, AI developer, or AI itself?
Originality: Can a work generated by AI be considered “original” in legal terms?
2. Detailed Case Laws on AI-Generated Art
Case 1: Naruto v. Slater (Monkey Selfie Case, 2018)
Facts:
A macaque monkey took a “selfie” using photographer David Slater’s camera.
Slater tried to claim copyright for the photo.
The claim was challenged by People for the Ethical Treatment of Animals (PETA) on behalf of the monkey.
Decision:
The court held that animals cannot hold copyright.
Copyright requires a human author.
Relevance to AI Art:
Sets a precedent that non-human creators cannot own copyright.
Suggests that AI-generated works may not automatically be copyrightable unless a human contributes sufficient creative input.
Case 2: Thaler v. Commissioner of Patents (DABUS AI, 2021, US & UK)
Facts:
Stephen Thaler listed an AI system called DABUS as the inventor of two patents.
The patents were rejected in the US, UK, and EU on the grounds that inventors must be human.
Decision:
Courts consistently ruled that only humans can be inventors under current patent law.
AI cannot be named as an inventor.
Relevance to AI Art:
Reinforces the idea that AI cannot be an author in copyright or patents.
Suggests ownership must belong to a human operator, programmer, or user.
Case 3: GitHub Copilot & Copyright Debate (Ongoing, 2023-2024)
Facts:
GitHub Copilot, powered by OpenAI Codex, generates code snippets based on public repositories.
Developers sued GitHub, claiming the AI copied copyrighted code without attribution.
Legal Issues:
Whether AI-generated outputs that closely resemble copyrighted material constitute infringement.
Whether the AI developer or user holds liability.
Implications for AI Art:
If AI art is generated using copyrighted works as training data, it may infringe existing copyrights.
Ownership and liability in AI-generated art are legally complex.
Case 4: Thaler v. Commissioner of Patents (DABUS AI, Australian Federal Court, 2019)
Facts:
Similar to the UK/US case but in Australia.
Thaler tried to patent an invention created by DABUS.
Decision:
Initially, the Australian court recognized AI as an inventor.
However, on appeal, the High Court reversed the decision: AI cannot hold patent rights; the patent must be assigned to a human.
Takeaways:
Shows the jurisdictional differences in treatment of AI.
Reinforces human-centric nature of intellectual property.
Case 5: Feist Publications, Inc. v. Rural Telephone Service Co., 499 U.S. 340 (1991)
Facts:
The case involved copyright protection for a phone directory.
The Supreme Court ruled that facts alone are not copyrightable, but creative selection or arrangement may be.
Relevance to AI Art:
Highlights the importance of originality and human creativity.
AI can generate content, but if it’s purely derivative or mechanical, it may not qualify for copyright.
Case 6: Thaler v. DABUS (European Patent Office, 2022)
Facts:
Thaler sought patent recognition for AI-generated inventions in Europe.
The EPO rejected the patent because an inventor must be a natural person.
Implications for AI Art:
Confirms that AI cannot be recognized as an author or inventor in Europe, reinforcing global legal consistency.
Human authorship remains the cornerstone for IP.
3. Key Takeaways from Cases
AI is not recognized as an author or inventor: Courts in multiple jurisdictions consistently rule that AI cannot hold IP rights.
Human involvement is crucial: Copyright or patent can only exist if a human contributes sufficient creativity or inventive input.
Derivative or AI-assisted works are complicated: If AI uses copyrighted works as input, it may create infringement risks.
Jurisdictional differences exist: Australia initially showed flexibility, while the US, UK, and EU are stricter about human authorship.
Future legislative change is likely: As AI-generated art becomes more common, laws may evolve to grant some form of ownership or sui generis rights to AI-assisted creations.
4. Summary Table
| Case | Year | Jurisdiction | Issue | Decision | Relevance to AI Art |
|---|---|---|---|---|---|
| Naruto v. Slater | 2018 | US | Animal copyright | Non-human cannot hold copyright | AI likely cannot hold copyright either |
| Thaler v. DABUS | 2021 | US/UK/EU | AI as inventor | AI not inventor | Human must be author/owner |
| GitHub Copilot | 2023 | US | AI-generated code | Ongoing | AI art from copyrighted data may infringe |
| Thaler v. DABUS | 2019 | Australia | AI as inventor | Initially yes, later reversed | Jurisdictional flexibility exists |
| Feist Publications v. Rural | 1991 | US | Originality | Only creative works copyrightable | AI works must show human creativity |
| EPO DABUS | 2022 | Europe | AI as inventor | Rejected | Confirms EU stance: humans only |
RELATED Blog
comments