USPTO Issues Revised 2025 Inventorship Guidance for AI-Assisted Inventions

The U.S. Patent and Trademark Office (USPTO) issued Revised Inventorship Guidance for AI-Assisted Inventions (Updated Guidance), providing a more streamlined and traditional framework for determining inventorship when AI tools are used in the inventive process. Effective November 28, 2025, the Updated Guidance rescinds previous USPTO guidance and reaffirms that inventorship remains grounded solely in human contribution and the longstanding conception standard for individual inventorship.

While the Updated Guidance does not modify any of the underlying legal doctrines for determining inventorship, it clarifies how AI-assisted innovation should be evaluated under existing doctrine and signals several areas where practitioners should exercise heightened scrutiny.

Key Takeaways

Pannu Factors Focused Solely on Natural Person Joint Inventorship

The Updated Guidance eliminates the 2024 USPTO that applied the Pannu joint-inventorship factors to questions of inventorship and AI tools in the innovation process. The Office now explicitly states that Pannu applies only to joint inventorship determinations between two or more natural persons and "[t]he fact that AI tools were used in the development process does not change the joint inventorship analysis among the human contributors." This position restores Pannu to its traditional context and removes any potential ambiguity introduced by the 2024 guidelines.

Importantly, the Updated Guidance does not provide new analysis or examples concerning the Pannu factors, or otherwise. In the absence of new commentary, the prior USPTO examples and discussions from 2024 remain the most detailed available, and practitioners may reasonably continue to rely on them when analyzing human-to-human joint inventorship scenarios.

The Significant Shift: Clarification of the Human Contribution Needed to Qualify as an Inventor

The Updated Guidance emphasizes that the controlling standard for inventorship of individual, natural persons—"the ability of the inventor to describe the invention with particularity"—is unchanged. Under traditional Federal Circuit precedent, a natural person qualifies as an inventor when they possess a definite and permanent idea of the invention and can describe it such that it may be reduced to practice without extensive research or experimentation.

Nonetheless, the Updated Guidance's framing introduces a nuanced implication—if a natural person may qualify as an inventor based on their ability to describe inventive concepts in a patent application, how much of the underlying concept can be generated by an AI tool, without impacting the natural person's inventorship? Using an edge case to examine the potential results of application of this policy, assume a natural person uses an AI tool to generate an output that includes information required to meet the Federal Circuit's conception standard not previously known to the natural person, i.e., an "invention." It would seem that the natural person could be named as an inventor simply by describing with particularity the inventive concept, even if the natural person did not separately conceive of critical aspects of the inventive concept without the outputs of the AI tool.

This potential broadening of the reach of "inventorship" for patent applications, also raises important questions under 35 U.S.C. § 102(f), which denies patentability if an applicant "did not invent the subject matter sought to be patented." Traditionally, § 102(f) has been applied in cases involving derivation from another natural person, not from a non-human, such as an AI-tool. The Updated Guidance does not discuss § 102(f) or offer any direction on how "inventing" should be interpreted in the context of AI-generated ideas. Similar questions related to the process of "inventing" may arise under 35 U.S.C. § 103, under which "patentability shall not be negated by the manner in which the invention was made" if the "manner" of inventing is with AI assistance. This almost certainly sets the stage for future evaluation of "edge cases," where courts and the USPTO will need to determine whether a human who merely recognizes, selects, or describes an AI-generated output has "invented" the subject matter.

This represents the most significant open issue arising from the Updated Guidance regarding inventorship. Although the Updated Guidance presents a number of open questions that will likely be addressed, it does not seem that the Updated Guidance will significantly modify the current approach to identifying potentially patentable subject matter and identifying inventorship contributions. As referenced in the Updated Guidance, the existing framework for inventorship already considers the use of various physical and computer tools by natural persons. But, perhaps one of the best practices that was commonplace prior to the implementation of the American Invents Act (AIA) in 2013, namely, documentation of the conception practices by potential inventors, will again be important to allow natural persons to readily identify their contributions to conception and reduction to practice in conjunction with AI-tools, especially with the possibility that the USPTO, the judicial branch, or the legislative branch may provide additional clarification in the future.

AI Systems Still Cannot Be Named as Inventors; Priority Claims Involving AI Inventors Still Barred

The Updated Guidance reaffirms existing law and doctrine:

• AI systems cannot be listed as inventors or joint inventors, and
• U.S. priority claims will not be accepted if the foreign or prior-filed application lists an AI system as an inventor.

These rules are not new, but the USPTO reiterates them given the increasing number of foreign applications that—under permissive foreign regimes—list AI systems as inventors. Applicants entering the U.S. national phase must continue to revise the Application Data Sheet to identify only natural persons as inventors.

Guidance Applies Equally to Utility, Design, and Plant Patents—Implications for AI-Generated Designs

Echoing prior USPTO statements, the Updated Guidance confirms that inventorship standards apply uniformly to utility, design, and plant patents. While this is not new, the reminder is significant because AI-assisted design generation is rapidly increasing.

Looking ahead, perhaps the first true AI-assisted inventorship analysis may arise in the design patent context and the use of AI-based tools. AI design tools can generate complex ornamental designs with minimal human input/prompts, raising questions such as:

• Does AI-output of design drawings satisfy the conception requirement in the context of design patents? Does selecting, refining, or approving an AI-generated design impact a determination of conception?
• What level of human input is required for a natural person to be considered the inventor of an AI-assisted design?

The Updated Guidance does not answer these questions, but by removing the 2024 Pannu analytical overlay for human-AI tool innovations, it may create greater flexibility for applicants seeking design protection for AI-generated physical articles.

Conclusion

The USPTO's Updated Guidance reflects a return to foundational inventorship principles while acknowledging the increasing role of AI in innovation. Although the guidance rejects any special inventorship framework for AI-assisted inventions, it leaves unresolved the most challenging conceptual question: what does it mean to "invent" when AI tools may generate or meaningfully shape the inventive concept?

As AI-assisted innovation accelerates, practitioners should expect further clarification—likely through case law—on § 102(f), conception, and design patent inventorship in the AI context.

Mauricio Uribe, Jeff Licitra, and Maria Anderson are recognized authorities in intellectual property law. As members of Davis Wright Tremaine's IP and patent practices, they provide valuable insights into complex legal and technological advancements, including the evolving role of AI in innovation. Stephanie Sanders is DWT's director of patent practice operations.

**

**

**