Ai Invention Disclosure Obligations.
📌 I. What Are “AI Invention Disclosure Obligations”?
In patent practice, inventors and patent applicants must accurately disclose the true inventors and provide a complete, enabling written description of the invention. When AI is involved — either as a tool in discovery or as a system that contributes to conception — questions arise such as:
Who is the inventor?
Does AI count as an inventor?
What must be disclosed about AI’s role?
Does use of AI affect patentability (novelty, obviousness, enablement)?
These are especially controversial because patents require human inventorship under current law in most jurisdictions.
📌 II. Core Legal Principles Underlying Disclosure Obligations
🔹 1. Inventorship Must Be Accurate
Patent laws (e.g., U.S. Patent Act, Indian Patents Act) require that only true inventors are named. Misnaming inventors can render a patent invalid.
🔹 2. Written Description & Enablement
The patent application must describe the invention fully so a person skilled in the art could replicate it — including any AI‑derived elements.
🔹 3. AI Is Generally Not Recognized as an Inventor
Courts and patent offices have so far held that only natural persons can be inventors.
🔹 4. Applicants Must Disclose AI’s Contribution
If an AI system significantly contributed to the conception, applicants must still explain what a human did to contribute to inventive steps.
📌 III. Detailed Case Discussions
Below are seven major cases (or case trends) that define how courts and patent offices treat AI and disclosure obligations.
📌 1. Thaler v. Vidal (U.S. Federal Circuit, 2022–2023)
Issue: Can an AI system be listed as an inventor?
Facts:
Dr. Stephen Thaler listed DABUS (an AI system) as the sole inventor on patent applications.
Ruling:
The Federal Circuit affirmed that:
Only natural persons can be inventors under U.S. patent law.
Therefore, listing an AI like DABUS as an inventor is improper.
The applications were invalid insofar as they named only AI.
Why It Matters for Disclosure:
This case clarifies disclosure obligations:
If AI contributes, the application must still identify human participants who contributed to the inventive concept.
Applicants cannot avoid disclosure by naming an AI as inventor.
Key Legal Takeaway:
Patent disclosure must reflect human intellectual contribution.
📌 2. Thaler v. Comptroller General (UK Supreme Court, 2021)
Issue: AI inventorship
Facts:
Similar to the U.S. case but decided by the UK Supreme Court: An application listed DABUS as an inventor.
Holding:
The Supreme Court held:
Only a natural person could be an inventor under the UK Patents Act.
The application was invalid where DABUS was listed as inventor.
Disclosure Impact:
This reinforces that, even if AI generates inventions, a human who directs, appreciates, or contributes to the invention must be identified and disclosed.
📌 3. European Patent Office (EPO) Practice — EPO Board of Appeal (2020–2023)
Issue: AI inventorship
Facts/Developments:
Several applications attempted to designate DABUS as inventor before the EPO.
Outcome:
The EPO consistently refused these designations because EU law recognizes only natural persons.
Disclosure Impact:
Applicants are obligated to:
Disclose details of AI’s role in the creation,
But name a human inventor, usually the person who conceived the problem, guided the AI, or interpreted its outputs.
📌 4. U.S. Patent Office (USPTO) Guidance on AI (2021)
Issue: How to disclose AI contributions
Guidance Summary:
The USPTO made clear that:
AI contributions must be disclosed in the specification if relevant.
Applicants must identify which elements were conceived by humans and which resulted from AI assistance.
Simply “merely using AI” does not necessarily create patentable subject matter — human contribution must be clear.
Why It Matters:
Patent applicants cannot bury the fact that AI was used. Omissions could lead to:
Invalid patent
Charges of inequitable conduct (withholding material information)
This guidance affects written description obligations.
📌 **5. U.S. District Court — Thaler v. Iancu (2019) (Predecessor to Vidal)
Issue: Inventorship & AI in utility patents
Summary:
Before the Federal Circuit, the District Court ruled that:
The Patent Act requires inventive entity to be natural persons.
AI cannot satisfy this requirement.
Disclosure Impact:
Even at this stage, courts instructed that applicants must describe what the human inventor actually did to conceive the invention.
📌 6. Williamson v. Citrix Online, LLC (2015) (Before AI Era, but Still Relevant)
Issue: Outsourced/Automated Contribution
Facts:
This case dealt with whether an automated demonstrator or third‑party programmer contributed to inventive concept.
Holding:
The Federal Circuit found that:
Individuals or systems that perform routine tasks (even automated) do not count as inventors.
Inventorship is about original intellectual contribution, not execution.
Why It Matters for AI:
When AI generates data or designs, the applicant must still show which humans contributed to the inventive idea, not just the mechanical work done by AI.
📌 7. Cases on Undisclosed Contributors & Inequitable Conduct
Although not always about AI, these cases illustrate the legal consequences of not disclosing material contributions.
★ Therasense, Inc. v. Becton, Dickinson & Co. (2011)
Failure to disclose material information to the patent office may render the patent unenforceable.
If AI substantially contributed to conception but was not properly detailed, a court might find inequitable conduct.
★ Kingsdown Medical Consultants v. Hollister (1988)
All individuals who made a significant contribution to conception must be listed.
AI Parallel:
If an AI system’s output is essential and a human cannot identify what exactly was contributed, that absence may jeopardize enforceability.
📌 IV. How These Cases Define Disclosure Obligations
🔹 1. Clearly Identify Human Inventors
Courts consistently hold that:
Only humans can be inventors.
AI cannot be listed on a patent.
🔹 2. Explain AI’s Role in the Description
Patent specifications should disclose:
How AI was used (e.g., generating a design, modeling, selecting materials)
Human decision points (what humans guided, selected, or modified)
🔹 3. Distinguish Human Conception from AI Output
Inventorship is about conception, not execution:
If the AI simply executed instructions, human inventors remain.
If AI generated novel concepts without human direction, applicants must show how humans were involved in conceiving and applying those concepts.
🔹 4. Full Enablement Includes AI‑Derived Inventions
The written description and enablement must allow a person skilled in the field to practice the full scope of the claimed invention — including AI‑derived methods.
📌 V. Practical Patent Filing Guidance (AI Context)
When drafting a patent involving AI:
| Step | What To Do |
|---|---|
| 1. Identify all contributors | List humans who directed or contributed to inventive ideas |
| 2. Describe AI usage | Explain role of AI tools or systems and human decisions |
| 3. Clarify inventive concept source | Show which elements are human‑conceived |
| 4. Avoid mislabeling AI as inventor | Only natural persons can be named |
| 5. Consider policy guidance | Follow USPTO/EPO guidance on AI disclosure |
📌 VI. Why This Matters for Innovation
Accurate disclosure:
Ensures valid patents
Avoids invalidity or unenforceability
Preserves clarity about AI’s role in innovation
Helps courts and examiners assess novelty and inventive contribution
📌 VII. Core Takeaways
AI cannot be an inventor — only natural persons.
Inventorship obligations still apply even with advanced AI tools.
Patent specifications must disclose AI’s involvement and human intellectual contribution.
Failure to disclose material information about AI’s role can invalidate a patent or lead to inequitable conduct findings.
RELATED Blog
comments