Criteria for Authorship in Innovation Disclosures, Patents and Copyright

This is a work in progress.
Last updated: February 11, 2026.
Note: This is practical lab guidance for transparency and consistency, not legal advice, but rather educational. Final determinations should be made with the tech transfer office and/or patent counsel.

Why this is different from a peer-review paper

A peer-review paper is a scholarly narrative. Authorship is governed mostly by academic norms and journal policies, and it can credit many kinds of contributions (idea, design, execution, analysis, writing, etc.).

Innovation disclosures, patents, and copyright are intellectual property with legal standards:

• Patents: the relevant concept is inventorship (not “authorship”). Inventorship depends on who conceived the claimed invention (often claim-by-claim), not who did the most work or wrote the most text.

• Copyright: the relevant concept is authorship of original expression fixed in a tangible medium (code, figures, text, artwork). Ideas alone are not copyrightable; expression is.

Bottom line: The set of people on a paper can (legitimately) differ from the set of patent inventors and from the set of copyright authors. Do not copy a paper author list into a disclosure or patent.

Quick comparison: papers vs patents vs copyright

Common scenarios

• Inventor but not paper author: conceived a key claimed feature early, but did not contribute to the paper narrative.

• Paper author but not inventor: executed experiments/simulations and analysis, but did not contribute to conception of claimed features.

• Copyright author but not inventor: implemented code that embodies an invention conceived by others.

Part A: Patents and Innovation Disclosures (Inventorship)

1) The core legal test: conception of the claimed invention

Inventorship is primarily about conception; who formed a definite and operative idea of the invention as reflected in the claims. Execution (“reduction to practice”) can support evidence, but it is not the core test.

Key points:

• Joint inventors can contribute unequally.

• Joint inventors do not need to work together or contribute to every claim.

• Inventorship can change as claims evolve.
  
  

2) Practical criteria: who should be listed in an innovation disclosure as a “potential inventor”

At the disclosure stage, it is usually better to be inclusive of plausible inventors, because the final claims can change and counsel can narrow the inventor list later. 

Strong signals someone is a (potential) inventor:

1. They proposed a specific technical solution that becomes a claim element (not just stating the problem).

2. They added a non-obvious feature/constraint at is necessary for the invention to work and is later claimed.

3. During development, they made an inventive change (not routine troubleshooting) that introduced a new claim limitation.

4. They contributed to conception over time on a complex invention (partial conception can still count if reflected in claims).
   
   

Usually NOT inventorship if they only:

• Followed instructions to run experiments/simulations/fabrication without contributing to conception.

• Performed routine testing, data collection, “button pushing,” standard pipeline runs, or ordinary debugging/optimization without proposing a new claimed feature.

• Provided materials, samples.

• Drafted or edited the disclosure/patent text without adding new technical substance.
  
  

3) Conception vs. development/execution: a “weights” heuristic (helpful, but not the law)

Patents are not awarded by point totals. A single crucial conceived element can make someone an inventor, while months of execution might not.

For internal discussions, it helps to be explicit:

• Conception of claim elements: dominant weight.

• Inventive refinement during development: counts only if it changes what is being claimed.

• Execution alone (reduction to practice): typically does not create inventorship.

• Writing: does not create inventorship unless it introduces new technical substance that becomes claimed.
  
  

4) Key differences from paper authorship (patents)

• There is no “author order” for inventorship.

• Inventorship is claim-by-claim; the inventor list can change as claims change.
  
  

Part A.1: AI-Assisted Inventions (latest USPTO guidance)

The USPTO’s Revised Inventorship Guidance for AI-Assisted Inventions (Nov 2025) emphasizes:

• The inventorship standard is unchanged.

• Only natural persons can be inventors; AI systems are tools, not inventors.
  
  

What counts in AI-assisted workflows:

• Humans can be inventors if they provide the inventive contribution (conception) reflected in the claims.

• Using AI does not automatically reduce inventorship; the question remains: did a person conceive the claimed invention?

• Human selection, modification, constraint-setting, and integration can matter when it amounts to conception of claim elements.

Lab practice recommendation: document human decision points (constraints, selection rationale, modifications, ablations, negative results, and when the “definite and operative” idea emerged).

Assigning “Contribution %” (internal transparency)

Important: “Contribution %” is an internal attribution tool (paper author-order discussions, internal records, optional contribution statements).
It does not determine patent inventorship, which is a legal, claim-based analysis. Do not use percentages to argue that someone “is” or “is not” an inventor.

When to use Contribution %

• Papers: to support author order discussions and generate a clear contributorship statement.

• Software/copyrightable outputs: to reflect who created the copyrightable expression (code, figures, text, UI).

• Patents/disclosures: only as a non-legal record of effort and conceptual contribution; inventorship is separate.
  
  

Step-by-step method (weighted role matrix)

1. Pick the output type (Paper, Patent/Disclosure, Software/Copyright) and use the default weights below as a starting point.

2. List contributors (include anyone who plausibly contributed).

3. Score each contributor on each role with a simple scale:

   • 0 = none

   • 1 = minor

   • 2 = supporting

   • 3 = substantial

   • 4 = leading

   • 5 = primary driver / owner

4. Compute weighted points for each person and normalize to 100% (formula below).

5. Sanity-check in a brief group review:

   • Does the result match who drove key decisions?

   • Are “invisible contributions” captured (mentoring, project admin, code review)?

   • Is any role overweighted relative to the output type?
     
     

Recommended lab process

1. Before any public disclosure (talks, posters, preprints, GitHub release): we need to chat about a possible submission of an innovation disclosure if there is any chance of patentable IP.

2. Treat inventorship as iterative: inventor lists can change as claims evolve.

References

1. USPTO Notice: Revised Inventorship Guidance for AI-Assisted Inventions (Federal Register, Nov 28, 2025)

2. USPTO Alert: Revised inventorship guidance for AI-assisted inventions (Nov 26, 2025)

3. Thaler v. Vidal, 21-2347 (Fed. Cir. Aug 5, 2022) (PDF)

4. Global Health Solutions LLC v. Selner, 23-2009 (Fed. Cir. Aug 26, 2025) (PDF)

5. 35 U.S.C. § 116 (Inventors)

6. USPTO MPEP § 2109.01 (Joint Inventorship)

7. Congressional Research Service: Artificial Intelligence and Patent Law (LSB11251)

8. U.S. Copyright Office Circular 1: Copyright Basics (PDF)

9. U.S. Copyright Office Policy Statement: Works Containing Material Generated by AI (Mar 16, 2023) (PDF)

10. Thaler v. Perlmutter, 23-5233 (D.C. Cir. Mar 18, 2025) (PDF)

11. CRediT (NISO): Contributor Role Taxonomy. https://credit.niso.org/ https://credit.niso.org/contributor-roles-defined/