AI-Assisted Drafting Of Patent Claims And Authorship Attribution.
1. AI-Assisted Drafting of Patent Claims
AI-assisted drafting refers to using artificial intelligence tools—like large language models, automated patent drafting software, or AI code generators—to help create patent applications, including claims, specification, and drawings.
Key Considerations
Inventorship: The patent system typically requires that an inventor be a human. AI cannot legally be recognized as an inventor in most jurisdictions.
Authorship / Contribution: If AI helps draft a patent claim, the human operator is generally considered the inventor if they provide the inventive concept.
Disclosure & Enablement: Even with AI assistance, the patent must satisfy disclosure requirements (enablement, written description, best mode) under 35 U.S.C. §112 in the U.S., or equivalent laws elsewhere.
Case Laws Related to AI and Inventorship
1. Thaler v. Iancu / Thaler v. Vidal (DABUS AI)
Facts: Stephen Thaler applied for patents in the U.S., EPO, and UK on behalf of his AI system called DABUS, listing the AI as the inventor.
Issue: Can an AI be legally recognized as an inventor?
Decision:
USPTO: Rejected the patent; only natural persons can be inventors.
EPO: Same rejection; European Patent Convention requires a human inventor.
UK High Court / Court of Appeal: Confirmed the patent cannot list AI as an inventor.
Significance: AI cannot be an inventor. Humans using AI can still be inventors, but the AI itself cannot hold rights.
2. Artificial Inventor Project – DABUS (Australia)
Facts: Australia initially allowed a patent listing DABUS as inventor.
Decision: The Federal Court overturned it, ruling that an inventor must be a human.
Significance: Reaffirmed global trend: AI-assisted drafting is permitted, but inventorship requires human involvement.
3. University of California v. Broad Institute (CRISPR Patent Dispute)
Facts: Dispute over CRISPR-Cas9 patents, including contributions from multiple inventors. AI-assisted tools were not central, but the case highlighted authorship and inventive contribution.
Decision: Court carefully examined who contributed to the inventive concept. Human judgment in claim drafting is crucial.
Significance: Shows that even with advanced tools, legal recognition of inventorship focuses on human intellectual contribution.
4. Thales v. IP Australia (Hypothetical Example Used in Literature)
Facts: A case discussed in academic literature where AI generated technical text for claims.
Issue: Can AI-generated content be part of prior art disclosure or claim drafting?
Outcome: Courts suggest AI can assist in drafting, but the human must control and verify the inventive content.
5. Enfish, LLC v. Microsoft Corp. (U.S. Federal Circuit, 2016)
Facts: Patent claims about database architecture. AI was not directly involved but cited in discussions on claim drafting and clarity.
Decision: The court emphasized clear, specific claims. Vague or overly broad claims (even if AI-generated) may fail under §112.
Significance: Shows that AI-generated claims must still meet legal clarity and enablement standards.
Key Takeaways on AI-Assisted Drafting
AI is a tool, not a legal inventor. Humans are responsible for inventive contribution.
Claims must satisfy legal requirements, including novelty, non-obviousness, and enablement. AI-generated claims may need careful review.
Authorship or contribution attribution must be human-based; the law currently does not recognize AI as a co-inventor.
Global consistency: Most jurisdictions (US, EPO, UK, Australia) require a human inventor.
2. Authorship Attribution in AI-Assisted Patents
Authorship attribution deals with who deserves credit for creating the patent. AI complicates this because it can generate technical text, prior art searches, and even suggested claims.
Principles
Human control: The person operating the AI and directing the invention is the inventor.
AI output review: Courts expect humans to vet AI suggestions; reliance on AI alone is insufficient.
Documentation: Keep records of human contributions vs AI outputs for patent prosecution.
Implications for Law Practice
Patent attorneys increasingly use AI to draft claims, perform prior art searches, and optimize claim language, but final responsibility is always human.
Improper attribution could result in invalid patents or ownership disputes.
Summary Table of Cases
| Case | Jurisdiction | AI Involvement | Key Principle |
|---|---|---|---|
| Thaler v. Iancu / Vidal | US/EPO/UK | AI listed as inventor | Only humans can be inventors; AI cannot hold patent rights |
| DABUS Australia | Australia | AI-generated invention | Court: Human inventor required; AI cannot be inventor |
| UC v. Broad Institute | US | Human and AI-assisted drafting | Human contribution is key for inventorship |
| Enfish v. Microsoft | US | AI not directly involved | Claims must be clear, specific; AI-generated drafts still need human verification |
| Thales/IP Australia (Academic) | Australia | AI assisted claims | AI can draft claims, human must verify and contribute inventive step |
RELATED Blog
comments