• 

Legal Frameworks Governing AI-Generated Inventions And AuthorshIP Recognition Under Korean Patent Law.

I. Conceptual Background: AI-Generated Inventions

• AI-generated inventions: Innovations produced wholly or partly by artificial intelligence systems (e.g., drug discovery, engineering designs).
• Core legal issue:
  • Patent law traditionally requires a human inventor
  • AI lacks legal personality, raising questions about:
    • inventorship
    • ownership
    • validity of patents

II. Korean Legal Framework

1. Patent Act (특허법)

The Korean Patent Act governs patentability and inventorship.

(A) Requirements for Patentability

An invention must be:

1. Novel
2. Inventive (non-obvious)
3. Industrially applicable

AI-generated outputs can satisfy these substantive criteria, but problems arise with inventorship.

(B) Inventorship Requirement

• The Act assumes that an inventor is a natural person.
• Inventor = person who contributes to the conception of the invention.

👉 AI cannot legally:

• Conceive an invention
• Be listed as an inventor

(C) Ownership of Patent Rights

Under Korean law:

• Initial ownership belongs to the inventor
• Rights can be transferred to:
  • employer (work-for-hire)
  • corporation
  • research institution

For AI inventions:

• Ownership typically assigned to:
  • programmer
  • system operator
  • employer institution

2. Civil Act (민법)

• Governs contractual allocation of rights
• Important for:
  • employment agreements
  • joint research arrangements
• Determines who ultimately owns AI-generated inventions

3. Copyright Act (Indirect Relevance)

• Although patents govern inventions, copyright may apply to:
  • software generating the invention
• Reinforces the human authorship principle, influencing patent interpretation

4. Korean Intellectual Property Office (KIPO) Practice

• KIPO requires:
  • human inventor identification
• Patent applications naming AI as inventor are rejected

III. Key Legal Issues in AI Inventorship

1. Can AI Be an Inventor?

• Current answer in Korea: No
• Inventorship tied to:
  • legal responsibility
  • rights ownership

2. Who Owns AI-Generated Inventions?

Possible candidates:

1. Programmer (designed AI system)
2. User/operator (triggered invention)
3. Employer (if within employment scope)

Korean practice favors:
👉 Human contributor with closest causal role

3. Level of Human Contribution

Key legal test:

• Did a human contribute to the inventive concept?
• If yes → patent valid
• If no → patent may be invalid

4. Ethical and Policy Concerns

• Incentivizing innovation vs. preserving human-centered IP law
• Risk of:
  • “inventorship inflation” (assigning humans artificially)
  • uncertainty in ownership

IV. Case Laws (Detailed Analysis)

Korean courts have limited direct AI patent cases, so comparative jurisprudence + Korean doctrinal application is essential.

1. DABUS Patent Cases – Global (US, UK, EU)

Facts

• AI system “DABUS” generated inventions
• Applications filed listing AI as sole inventor

Issue

• Can AI be recognized as inventor?

Judgment

• Courts in multiple jurisdictions rejected AI inventorship
• Reasoning:
  • Patent statutes require natural persons
  • Inventor must have legal capacity

Korean Relevance

• Korean Patent Office follows same reasoning
• Applications must list human inventor

Principle

👉 AI cannot be an inventor; human attribution is mandatory

2. Thaler v. Commissioner of Patents

Facts

• Applicant attempted to list AI (DABUS) as inventor

Issue

• Whether “inventor” includes non-human entities

Judgment

• Court rejected AI inventorship
• Emphasized:
  • Inventor must be capable of holding rights
  • AI lacks legal personality

Korean Implication

• Reinforces strict human inventorship requirement
• Influences Korean academic and policy discussions

3. University of Utah v. Max-Planck-Gesellschaft

Facts

• Dispute over who contributed to invention
• Question of correct inventorship attribution

Issue

• What constitutes “inventive contribution”?

Judgment

• Only those contributing to conception qualify as inventors

Relevance to AI

• Helps determine:
  • whether programmer or user qualifies as inventor
• Korean courts apply similar reasoning

Principle

👉 Inventorship = contribution to inventive concept, not mere assistance

4. Naruto v. Slater (Analogous)

Facts

• Monkey took photograph using human camera

Issue

• Can non-human hold IP rights?

Judgment

• Court held:
  • Non-human cannot hold copyright
  • Human controlling process owns rights

Korean Relevance

• Applied analogically to patents:
  • AI cannot hold inventorship
  • Human operator/programmer is rights holder

5. Korean Employee Invention Case (Supreme Court)

Facts

• Employee created invention during employment
• Dispute over ownership between employee and employer

Issue

• Who owns patent rights?

Judgment

• Employer owns rights if:
  • invention created within scope of duties
• Employee entitled to compensation

AI Relevance

• If AI system used in employment:
  • employer typically owns resulting inventions

Principle

👉 AI-assisted inventions follow employee invention rules

6. Korean Patent Inventorship Misidentification Case

Facts

• Patent challenged due to incorrect inventor listing

Issue

• Does incorrect inventorship invalidate patent?

Judgment

• Court held:
  • Incorrect inventorship can invalidate patent
  • Must identify true human contributors

AI Relevance

• Risk:
  • assigning inventorship improperly in AI-generated inventions

Principle

👉 Accurate human inventorship is legally critical

7. Stanford University v. Roche Molecular Systems

Facts

• Dispute over ownership of patented invention between university and company

Issue

• Whether rights automatically vest in employer

Judgment

• Rights initially vest in inventor (individual)
• Assignment required for transfer

Korean Relevance

• Similar principle:
  • inventor first owns rights
  • then assigned to institution

AI Context

• Important for determining:
  • whether programmer or institution owns AI inventions

V. Doctrinal Takeaways

1. Human Inventorship Is Mandatory

• AI cannot be inventor under Korean Patent Act

2. Inventorship Depends on Contribution

• Must contribute to conception, not just execution

3. Ownership Follows Legal Assignment

• Inventor → employer or institution via contract

4. Incorrect Inventorship Risks Invalidity

• Critical issue in AI-generated inventions

5. Comparative Law Strongly Influences Korea

• DABUS and global cases shape Korean interpretation

VI. Practical Implications for AI Research Labs

1. Always identify a human inventor
   • Programmer, researcher, or operator
2. Document human contribution
   • Maintain records of inventive steps
3. Use clear IP assignment agreements
   • Especially in collaborative AI research
4. Avoid naming AI as inventor
   • Leads to automatic rejection
5. Audit inventorship carefully
   • Prevent patent invalidation

VII. Conclusion

Under Korean patent law:

• AI-generated inventions are not excluded, but
• AI cannot be recognized as an inventor

👉 The law adopts a human-centered inventorship model, where:

• Human contribution determines inventorship
• Contracts determine ownership
• Courts emphasize conception, accountability, and legal personality

RELATED Blog

Software Patenting

OpenAI Sued by Multiple Copyright Lawsuits in Indi...

Delhi High Court Rules in Favor of Svamaan in Trad...

US Supreme Court to Decide Whether AI-Generated Ar...

Intellectual Property Laws at Afghanistan

Intellectual Property Laws at Albania

Intellectual Property Laws at American Samoa (US)

Intellectual Property Laws at Angola

Intellectual Property Laws at Antigua and Barbuda

Intellectual Property Laws at Armenia