AI-Assisted Invention Patent Disputes.
I. Understanding AI-Assisted Invention in Patent Law
1. What Is an AI-Assisted Invention?
An AI-assisted invention is one where:
A human uses AI as a tool to help generate ideas, designs, or solutions, or
AI autonomously generates an invention with minimal or no human involvement.
Patent law globally is based on the assumption that an inventor must be a natural person, capable of:
Conception
Intent
Legal accountability
This creates tension when AI systems perform inventive tasks.
II. Core Legal Issues in AI Patent Disputes
Inventorship – Can AI be named as an inventor?
Ownership – Who owns an AI-generated invention?
Patentability – Does AI-generated output meet novelty and non-obviousness standards?
Disclosure – How much must be disclosed about AI’s role?
Liability – Who is responsible for infringement or errors?
III. Major Case Laws on AI-Assisted Invention
CASE 1: Thaler v. Commissioner of Patents (Australia, 2021–2022)
Facts
Stephen Thaler developed an AI system called DABUS
DABUS generated inventions without human intervention
Thaler filed patent applications naming DABUS as the sole inventor
Legal Question
Can an AI system be legally recognized as an inventor?
Decision
Trial Court (2021): Allowed AI as inventor
Full Federal Court (2022): Reversed decision
Reasoning
Patent statutes consistently use terms like “person”
Inventorship requires legal personality
Ownership rights cannot vest from a non-person
Significance
Temporarily made Australia the first country to accept AI inventorship
Final ruling reaffirmed human-only inventorship
Highlighted legislative gaps
CASE 2: Thaler v. United States Patent and Trademark Office (USA, 2022)
Facts
Same DABUS inventions filed with USPTO
AI named as inventor
Legal Question
Does US patent law allow non-human inventors?
Decision
Patent application rejected; upheld by federal court
Court’s Reasoning
US Patent Act defines inventor as an “individual”
Supreme Court precedents define “individual” as a natural person
Policy concerns: AI cannot take oaths, assign rights, or be cross-examined
Importance
Established clear precedent in the US
Reinforced that AI is a tool, not an inventor
Frequently cited in later AI-patent disputes
CASE 3: Thaler v. Comptroller-General of Patents (United Kingdom, 2023 – Supreme Court)
Facts
Thaler named DABUS as inventor
Claimed ownership as the AI’s owner
Legal Question
Is ownership of AI sufficient to claim inventorship rights?
Supreme Court Decision
Unanimous rejection
Legal Analysis
UK Patents Act requires an inventor to be a person
Ownership of a machine ≠ ownership of inventions
No legal mechanism exists for AI to transfer rights
Impact
Strongest judicial rejection of AI inventorship
Confirmed inventorship and ownership are inseparable
Influential globally
CASE 4: European Patent Office (EPO) – DABUS Applications (2020–2021)
Facts
Patent applications filed naming DABUS as inventor
Applicant argued AI autonomy
Decision
Applications refused
Reasoning
Inventor must have:
Legal capacity
Ability to exercise rights
AI lacks moral and legal personhood
Disclosure rules require human identification
Significance
First major administrative rejection
Set the tone for EU policy
Emphasized procedural impossibility of AI inventorship
CASE 5: In re Application of Artificial Intelligence Inventorship (China, CNIPA Practice)
Context (Not a single case, but authoritative rulings)
China rejected AI-only inventorship but accepted AI-assisted inventions
Key Position
AI may contribute to inventive steps
A human who controls, selects, or validates output is the inventor
Legal Importance
More flexible than Western jurisdictions
Focuses on human control and contribution
Encourages AI innovation without changing inventorship rules
CASE 6: South African DABUS Patent Grant (2021)
Facts
South African patent office granted a patent naming DABUS as inventor
Why This Happened
South Africa has a non-substantive examination system
No deep legal analysis conducted
Legal Value
Not a binding precedent
Frequently misunderstood as recognition of AI inventorship
Later clarified as procedural, not doctrinal
CASE 7: Recent AI-Assisted Pharma Patent Disputes (US & EU)
Context
AI used to:
Identify drug candidates
Predict molecular structures
Legal Issue
Opponents argued:
Lack of human inventive step
Obviousness due to algorithmic generation
Rulings
Courts upheld patents where:
Humans selected final candidates
AI used as an advanced research tool
Principle Established
AI does not negate inventiveness if human judgment is involved
IV. Comparative Legal Position (Summary)
| Jurisdiction | AI as Inventor | AI-Assisted Invention |
|---|---|---|
| USA | ❌ No | ✅ Yes |
| UK | ❌ No | ✅ Yes |
| EU | ❌ No | ✅ Yes |
| China | ❌ No | ✅ Yes (Flexible) |
| Australia | ❌ No | ✅ Yes |
| South Africa | ⚠️ Procedural anomaly | Unclear |
V. Key Legal Principles Emerging
AI cannot be an inventor
Human involvement is mandatory
Ownership flows from inventorship
AI is treated as a sophisticated tool
Disclosure of AI use may become mandatory
VI. Future Outlook
Legislatures may introduce:
AI-specific inventorship categories
AI contribution disclosure requirements
Courts currently prefer evolution, not revolution
Ethical and accountability concerns dominate legal reasoning
VII. Conclusion
AI-assisted invention disputes reveal a fundamental truth:
Patent law protects human creativity, not machine autonomy.
Until laws change, humans must remain at the center of inventorship, regardless of how powerful AI becomes.
RELATED Blog
comments