USPTO Issues Revised Inventorship Guidance for AI-Assisted Inventions

The US Patent and Trademark Office recently issued revised guidance clarifying that only natural persons may be named as inventors on patent applications involving AI-assisted inventions. The guidance rescinds earlier approaches that applied joint-inventorship analyses to scenarios involving human-AI collaboration.

The guidance states that traditional conception standards, not new or modified tests, govern inventorship, even where AI tools play a significant role.

For inventors and companies, the guidance reinforces the need to document human contribution, structure internal innovation processes to clearly identify inventors, and avoid naming AI systems in application materials.

BACKGROUND

The revised guidance, released on November 26, 2025, replaces the US Patent and Trademark Office’s (USPTO’s) February 2024 inventorship guidance and removes confusion about when joint-inventor principles, such as the Pannu factors, apply. Because AI systems are not persons and cannot legally be inventors, joint-inventorship analysis applies only when multiple human contributors are involved.

The USPTO emphasizes that AI tools—whether generative, analytical, or model-based—are treated the same as laboratory equipment, software, or research databases: meaning, they may assist in the creation of inventions, but they cannot conceive such inventions.

KEY TAKEAWAYS FOR INVENTORS AND COMPANIES

AI May Assist but Cannot Be Credited as an Inventor

Any application listing an AI system as an inventor or joint inventor will be rejected under 35 USC §§ 101 and 115.

Practical advice:

• Ensure all invention disclosure forms require identification of only natural-person contributors.
• Update internal policies to explicitly prohibit filing patents for AI-conceived inventions and naming AI tools as inventors on any patent filings.
• Maintain version control for AI-generated materials to track how human inventors conceived of the invention, and refined, selected, and/or integrated AI outputs (for example, including the selection of inputs provided to AI tools and any configuration or training of the AI tools).

Document Human-Driven Conception

Conception requires the human inventor to form a “definite and permanent idea” of the complete and operable invention. To qualify, the human inventor must have a specific solution, not only a general research goal.

Practical advice:

• Require inventors to document their thought conception, including processes, problem framing, and design choices before running or integrating AI tools.
• When AI proposes multiple outputs, document why the inventor selected or modified a particular output.
• Use timestamps and collaborative tools to show the inventor’s role throughout development.

Joint-Inventorship Principles Still Apply for Humans

If several people contribute materially to conception or reduction to practice, the traditional Pannu factors govern inventorship among the human contributors. AI does not alter this analysis.

Practical advice:

• Implement structured inventor-interview checklists during patent preparation.
• Document each person’s contribution to specific limitations of the claims or features of the invention.
• Resolve inventorship questions early before drafting begins to avoid later correction challenges.

Patent Filings Must Reflect Human Inventors Even for Foreign Priority Claims

The USPTO will not accept priority claims to foreign applications listing an AI system as an inventor. US applications must list only natural persons, even if foreign law allows non-human inventors.

Practical advice:

• Review foreign counsel’s filings to ensure no AI inventors are listed.
• If there is an error in incorrectly naming an AI inventor, correct foreign priority documents before filing in the United States, if possible.
• Ensure international research and development (R&D) teams understand US inventorship requirements.

Guidance Applies to Utility, Design, and Plant Patents

Under the new guidance, the same conception standards govern all three categories of patent protection. Inventors must have contributed to the creation of the claimed subject matter.

Practical advice:

• For AI-generated designs, document the inventor’s conception, including selections, configurations, and/or modifications that reflect a definite and permanent idea.
• For plant innovations assisted by AI (for example, predictive phenotyping), ensure that the inventor’s role in creation and asexual reproduction is clear.

STRATEGIC INSIGHT

The USPTO’s reiteration of traditional conception doctrine creates several operational implications for organizations adopting AI-assisted R&D:

• Governance and Compliance: Companies should treat AI-assisted inventorship risks as part of broader AI governance frameworks. Policies should require both human-in-the-loop conception, validation and documentation of human conception and judgment.
• Evidence Preservation: In future disputes, particularly during post-grant challenges, companies may need to prove human conception with contemporaneous documentation. Strong documentation practices can provide defensive value.
• Invention Harvesting Optimization: Organizations should reassess their invention disclosure procedures to ensure that R&D teams articulate human-driven conception, not just AI outputs. This may require training inventors on how to properly describe their contributions.
• Portfolio Strategy: This guidance may affect global filing strategies. Since some foreign jurisdictions may evolve toward broader recognition of AI invention, companies need consistent internal standards to avoid conflicts in priority chains.
• Risk of Misrepresenting Inventorship: Incorrect inventorship can render patents unenforceable or invalid. Companies should implement internal audits of inventorship determinations, particularly for AI-assisted inventions.

LOOKING AHEAD

The USPTO’s guidance signals an ongoing focus on AI-assisted innovation and increased scrutiny of inventorship assertions as technologies evolve. Applicants should expect continued guidance and potential future rulemaking as AI tools become more sophisticated.

Companies leveraging AI in R&D should proactively implement internal controls, training, and documentation protocols to align with USPTO expectations. Staying ahead of these standards will reduce prosecution risk and strengthen long-term enforceability.