The United States Patent and Trademark Office (USPTO) has issued updated examination guidance regarding inventorship in applications related to artificial intelligence (AI) (the “New Guidance”). This document revokes and replaces the February 13, 2024 guidance that clarifies how inventorship should be determined when AI is used in the invention process. The new guidance is Pannu For this purpose, the test focused on the co-inventor issue and instead focused on ideation. This action is another step by the new USPTO leadership to strengthen the patent system. It remains to be seen whether the courts will agree with this approach. Even if some patents are granted by the USPTO under this guidance, they may be held invalid by a court. This will still largely depend on the facts. Below is a detailed breakdown of the key changes across utility, design and plant applications and their practical impact on patent strategy.
Big Reset: Reversal of previous AI inventorship guidance
The USPTO has expressly rescinded the February 13, 2024 Inventorship Guidance for AI-Assisted Inventions in its entirety and rescinded the application of the Pannu co-inventor element to AI-assisted inventions as part of the general inventorship framework. The Office emphasizes that pannu was and remains the principle governing joint inventorship between two or more natural persons. This does not apply if the only other “participant” is an AI system, which by definition is not human. In short, the previous approach was off the table, and the agency refocused its analysis on traditional legal principles regarding ideation and inventorship.
No “co-inventor” question required if AI is the only other entity
The new guidance squarely states that a co-inventor investigation will not arise if a single natural person uses an AI during development, as an AI system is not a “person” and therefore cannot be a co-inventor. This resolves previous confusion about whether to perform co-inventorship analysis in a single human versus AI scenario. The answer is no. The existence of an AI tool does not by itself trigger multi-inventor analysis.
A single, unified standard: Traditional concepts govern AI-assisted work
The USPTO emphasizes that the legal standards for inventorship are the same for all inventions, and there are no separate or modified standards for AI-assisted inventions. Based on Federal Circuit precedent, only natural persons can be inventors. Therefore, AI, sophisticated or not, cannot be named as an inventor or co-inventor. The touchstone of being an inventor remains “conception,” defined as the formation in an inventor’s mind of a clear and lasting idea of a complete and working invention. Visioning requires concrete, well-established ideas and specific solutions to the problem at hand. General goals and research plans are insufficient.
Ideas require commitment and a sense of ownership.
The new guidance reiterates that inventorship is a highly fact-driven inquiry that focuses on whether the inventor has knowledge of all limitations of the claimed invention and could practice the invention using ordinary skill alone, without extensive research or experimentation. Analysis enables the ability to specifically explain the invention. Without such an explanation, the inventor cannot objectively demonstrate that he has a complete mental image. This framework reinforces the benefits of robustly and simultaneously documenting an inventor’s mental steps and the specific claims that result from them, especially when AI is used to generate or refine the input that informs the claimed features.
AI as a tool: presumptions, naming practices, and rejection.
As a practical matter, the USPTO maintains a presumption that the inventor listed on the application data sheet or oath/declaration is the actual inventor. However, the Patent Office has directed that rejection under 35 USC §§ 101 and 115 (or other appropriate action) apply to all claims in applications that list an AI system or other non-natural person as an inventor or co-inventor. Conceptually, AI systems, including generative AI and computational models, are “equipment” used by human inventors, similar to laboratory equipment, software, and databases. Inventors can utilize the “services, ideas, and assistance of others” without those sources becoming co-inventors. The same principle applies to AI systems. If a natural person is involved, the only question is whether that person conceived the invention according to traditional criteria of conception.
Multiple human contributors: Co-inventorship still uses Pannu
If more than one natural person is involved in the creation of an invention with the assistance of AI, traditional joint inventorship principles apply. Pannu Factors applied to determine whether each person qualifies as a joint inventor. Each person named as an inventor must (1) contribute in some significant way to the conception or practical application; (2) makes an insignificant contribution in terms of quality compared to the complete invention; (3) Go beyond simply explaining well-known concepts or cutting-edge technology. Importantly, the mere involvement of AI tools does not change the analysis of co-inventorship between human contributors. Practitioners should evaluate and document each person’s inventive contribution to the claim as usual.
Beyond practicality: also covers design and plant patents
The new guidance confirms that the same inventor search applies to design and utility patents. For plant patents, statute and Federal Circuit precedent require that the inventor contributed to the creation of the plant (rather than merely discovering and asexually copying it). The USPTO has made clear that these principles apply similarly when AI assists in the design and development of plant varieties. As a result, applicants will need to treat AI involvement in all types of patents like any other tool and demonstrate human inspiration and contributions consistent with applicable law and case law.
Claiming priority and benefit: identity of inventor must match
For applications and patents claiming interest or priority (U.S. or foreign) under 35 USC §§ 119, 120, 121, 365, or 386, the named inventors must be the same or have at least one designated joint inventor in common and must be natural persons. Priority claims to foreign applications that designate an AI tool as the sole inventor will not be accepted. If the foreign application names both natural persons and non-natural persons (such as an AI) as co-inventors, the U.S. application must list only natural persons, including at least one person in common with the foreign application. The same approach applies to national phase applications under 35 USC § 371. List the natural person’s name as the inventor on the application data sheet that accompanies the initial U.S. filing.
Practical implications and action items for patent strategy
- Thoroughly document human concepts. Prior to or concurrently with the use of the AI, maintain a detailed inventor notebook (electronic or paper) that describes in the inventor’s own words a clear and lasting idea of the complete and valid invention, particularly the claim limitations. Capture prompts, inputs, iterations, and human analysis to distill AI output into specific claims language or design features. This helps establish the ownership and distinctiveness of the inventor.
- When possible, treat AI output like experimental results. If AI has “conceived” part of the “invention” to be claimed, carefully analyze whether the human idea is sufficient. Position AI as a means to provide data and suggestions. The decisive step is not simply to receive AI-generated content, but for humans to form complete, practical inventions in their minds: specific solutions and billable capabilities.
- Coordinate cross-border filing strategies. For jurisdictions that allow non-human inventors, plan ahead to avoid conflicts with U.S. requirements. Try to keep at least one natural person inventor in common in the priority chain and exclude non-natural persons from inventorship in U.S. applications. Coordinate with foreign attorneys on naming conventions to maintain U.S. priority.
- Prepare for examiner scrutiny. In applications that discuss AI extensively, it is anticipated that examiners may question inventorship. Provide documentation of human conception and specific claim limitations conceived by the human inventor, as well as appropriate inventor declarations.
What hasn’t changed
- Despite evolving capabilities, AI cannot be named as an inventor in the United States, and the standards for inventorship have not changed. In other words, human ideas are central.
- Traditional inventor principles still apply. Conception, co-inventorship, and priority requirements continue to be based on statutory and Federal Circuit principles. AI is treated as a tool, not as a co-inventor.
- Assertion drives analysis. Whether a single inventor or many inventors, the question is who (or who) came up with the limitations of the claimed invention? The existence of AI does not change the claims-centric nature of inventorship evaluation.
conclusion
The USPTO’s new guidance simplifies and clarifies the path forward. There is no AI-specific inventorship test, AI cannot be named as an inventor, and traditional concept standards apply. When there is a single inventor, AI-assisted cases focus on documenting the human idea in detail. For scenarios with multiple inventors, continue to apply Pannu Among human contributors. Extend these principles across utility, design, and plant applications to ensure that priority claims are consistent with natural person inventorship and avoid fatal defects. With careful documentation and argument-focused analysis, AI can be a powerful vehicle for innovation while maintaining clear and defensible human inventorship.
