As artificial intelligence takes on a larger role in research and development, one question keeps surfacing for inventors and innovation teams: if an AI system generates a breakthrough idea, who gets named on the patent? And more fundamentally, can AI be an inventor at all?

Under U.S. patent law, the answer right now is no. Only a human being can be named as an inventor. But the full picture is more nuanced than that one-word answer suggests, and understanding it matters practically for any organization that uses AI tools in its innovation process.

At Goldstein Patent Law, we help inventors and companies navigate exactly these questions. Here’s where the law stands today, how it got here, and what it means for protecting AI-assisted inventions.

can ai be an inventor

The Clear Legal Answer Under U.S. Patent Law

The Patent Act defines an inventor as “the individual” who invents or discovers the subject matter of the patent. The word “individual,” as courts have interpreted it, means a natural person, a human being. That reading is reinforced throughout the statute by the use of personal pronouns: “himself,” “herself,” “themselves.”

On that basis, an AI system cannot be named as an inventor on a U.S. patent application. This is not a policy preference or an agency guideline. It is a settled legal holding confirmed at the Federal Circuit level and left standing by the Supreme Court. Any application that names a machine or AI system as an inventor will be rejected for failing to identify a valid inventor.

The Case That Settled It: Thaler v. Vidal

The legal test case on AI inventorship came from computer scientist Stephen Thaler, who filed two patent applications in 2019 naming his AI system, DABUS (Device for the Autonomous Bootstrapping of Unified Sentience), as the sole inventor. The inventions at issue were a fractal beverage container and a neural flame device designed for search-and-rescue operations. Thaler argued that DABUS had generated these inventions autonomously, without meaningful human creative input, and that he was simply the owner rather than the inventor.

The USPTO rejected both applications for failing to name a valid inventor. Thaler sought judicial review, and the U.S. District Court for the Eastern District of Virginia granted summary judgment for the USPTO. On appeal, the Federal Circuit affirmed in Thaler v. Vidal, 43 F.4th 1207 (Fed. Cir. 2022).

The court’s reasoning was a straightforward statutory-text analysis: the Patent Act defines “inventor” as “the individual” who invents, and elsewhere consistently uses personal pronouns to describe inventors. The court also noted that the Supreme Court has held that, unless Congress indicated otherwise, the word “individual” in statutes refers to a human being, and that requiring an inventor to be a human being is consistent with precedent holding that neither corporations nor sovereigns can be inventors because they are not natural persons.

The Supreme Court denied certiorari on April 24, 2023, without comment, leaving the Federal Circuit’s human-inventor requirement as governing U.S. law. The question of whether AI can be an inventor in the U.S. is, for now, closed.

How Other Countries Have Answered the Same Question

Thaler pursued parallel patent applications in multiple countries with the same DABUS applications, testing the question of AI inventorship internationally. The results were largely consistent with the U.S. outcome.

The European Patent Office rejected the application. The UK Supreme Court unanimously rejected AI inventorship in December 2023. Australia’s Federal Court initially allowed AI inventorship, but the Full Court reversed on appeal. Canada, Japan, South Korea, and Germany all rejected AI inventorship.

South Africa was the sole exception, granting a patent listing DABUS as inventor. Legal observers have noted this was likely a function of South Africa’s largely administrative patent registration system rather than a substantive legal decision, and it has not been followed elsewhere.

The international consensus reflects the same underlying logic as the U.S. ruling: patent statutes around the world were written with human inventors in mind, and changing that requires legislative action, not court interpretation.

The USPTO’s Evolving Guidance on AI-Assisted Inventions

While Thaler v. Vidal answered whether AI can be an inventor cleanly, it left open the more practical question that most companies actually face: what happens when a human uses AI tools extensively in the inventive process? Does heavy reliance on AI affect the human’s claim to inventorship?

The USPTO has issued two rounds of guidance addressing this. The first, in February 2024, tried to adapt the Pannu factors (a joint-inventorship framework originally designed to evaluate contributions among multiple human co-inventors) to the AI-human context, requiring evidence of a human’s “significant contribution” to distinguish genuine invention from mere prompting of an AI tool.

That guidance drew criticism for overcomplicating inventorship analysis and potentially chilling AI-assisted innovation by imposing undue hurdles on human inventors. The revised guidance, issued on November 28, 2025, under the Trump administration’s Executive Order on AI, fully rescinds the February 2024 guidance and replaces it with a streamlined framework that reaffirms that only natural persons can be inventors, treats AI systems as mere instruments analogous to other inventor support tools such as software and laboratory equipment, and applies the traditional legal standard for inventorship to all patent applications regardless of AI involvement.

The 2025 guidance clarifies that the Pannu factors apply only to collaborations between multiple natural persons, not to the question of whether a human inventor using an AI tool has met the inventorship standard.

What “Conception” Means When AI Is Involved

The traditional legal test for inventorship centers on conception: did the human form a definite and permanent idea of the complete and operative invention? A human inventor must form in their mind a definite and permanent idea of the complete and operative invention. That standard, under the 2025 guidance, applies without modification when AI tools are part of the process.

The USPTO guidance finds that humans may use AI tools when developing new ideas without forfeiting human inventorship status in patent applications. The guidance likens AI systems to traditional equipment that assists in the development process, with humans deploying those tools still being the ones who conceived the idea.

This is significant for anyone asking whether AI can be an inventor in any practical sense: the answer remains no, and using AI heavily in your process does not disqualify you as the human inventor, provided you exercised genuine creative judgment in directing, interpreting, and applying the AI’s outputs.

What matters is the nature of the human’s role. Recognizing that an AI-generated output represents a useful solution, selecting it from among alternatives, and understanding why it works are all forms of human conception. Passively accepting whatever an AI produces without meaningful creative evaluation is where the inventorship claim becomes vulnerable.

can ai be an inventor

The Open Question: What If AI Does Most of the Conceptual Work?

The 2025 guidance is clear in principle but leaves a genuinely hard question at the edges. 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.

As AI systems become more capable, the line between “tool that assists human conception” and “system that does the conceiving” will become harder to draw. The law’s current answer, that this is a question about the human’s creative role rather than the AI’s capabilities, is workable for now. But it is likely to face pressure as AI systems generate increasingly complex and autonomous outputs in drug discovery, materials science, and other innovation-intensive fields.

For companies operating at the frontier of AI-assisted R&D, the practical takeaway is to build documentation practices now that clearly establish the human conceptual contributions in every inventive step, before any legal challenge forces the question.

What This Means in Practice for R&D Teams

Whether AI can be an inventor as a legal matter is settled. But the operational question of how to manage inventorship correctly when AI tools are embedded in your R&D process is very much alive. A few practical implications:

Document human decision-making at every stage. The conception standard requires evidence of human creative judgment. Detailed records of which outputs were selected and why, what problems the human was trying to solve, and what made one AI-generated result preferable to another all support inventorship claims in the event of a challenge.

Do not list AI systems in application materials. Patent applications that name a machine on an application data sheet, an inventor’s oath or declaration, or a substitute statement as either an inventor or joint inventor will be considered by the USPTO to have improper inventorship. Even where AI played a substantial role, naming it as an inventor will result in rejection.

Evaluate joint inventorship carefully. When multiple humans collaborate on an AI-assisted invention, each person claiming inventorship must still have contributed to the conception of at least one claim. The Pannu factors continue to apply to this analysis among human co-inventors, and AI involvement does not simplify that evaluation.

Think beyond inventorship. Patent protection for AI-assisted inventions is one layer of a broader IP strategy. Trade secret protection for training data and model weights, and copyright for original software code, can complement patent rights. Our overview of the difference between patent and trademark is a useful starting point for thinking about IP coverage more broadly, and our post on patenting AI models covers the subject matter eligibility questions that apply once inventorship is established.

can ai be an inventor

Building a Strong AI-Assisted Patent Application

Inventorship is one piece of a broader set of considerations for anyone protecting AI-assisted innovation. Subject matter eligibility under § 101, non-obviousness under § 103, and written description requirements all present distinct challenges in this space. Our post on non-obviousness in patent law covers how examiners evaluate that standard in detail, including the secondary considerations that can be decisive when prior art is dense.

The patent application process for AI-assisted inventions follows the same fundamental steps as any utility patent, but the inventorship documentation, claim drafting, and § 101 analysis all require additional care. Starting with a provisional application can lock in your filing date while you refine both the claims and the inventorship record.

Frequently Asked Questions

Can AI be an inventor on a U.S. patent?

No. Under current U.S. law, as confirmed by Thaler v. Vidal (Fed. Cir. 2022) and left standing by the Supreme Court, only natural persons (humans) can be named as inventors on a patent. Any application naming an AI system as an inventor will be rejected.

Does using AI tools to develop an invention affect my status as the inventor?

No, provided you exercised genuine creative judgment in the process. The USPTO’s November 2025 revised guidance confirms that AI tools are treated like laboratory equipment or software: they assist the inventive process but do not affect the human inventor’s status. The key is that a human must have conceived the invention in the legal sense, forming a definite idea of the complete and operative solution.

What if the AI system did most of the creative work and the human only recognized that the output was useful?

This is the genuinely open question at the edges of current guidance. The USPTO and courts have not yet drawn a clear line between meaningful human conception and merely selecting from AI outputs. Companies in this situation should document human decision-making carefully and consult a patent attorney before filing.

Has any country allowed AI to be named as an inventor?

South Africa granted a patent listing DABUS as inventor, but this is widely attributed to the administrative nature of South Africa’s patent system rather than a substantive legal holding. Every major patent jurisdiction, including the European Patent Office, the United Kingdom, Australia (on appeal), Canada, Japan, South Korea, and Germany, has rejected AI inventorship.

What happens to inventions that AI genuinely created without meaningful human input?

Under current law, they may not be patentable in the U.S. at all, since a valid inventor must be named. This is one of the most significant open policy questions in IP law, and it is one that Congress, not the courts, will likely need to resolve if AI autonomy continues to advance.

Protecting Your AI-Assisted Innovation

The legal answer to whether AI can be an inventor is clear today. But the practical questions around inventorship documentation, claim strategy, and IP coverage for AI-assisted work require careful, case-specific analysis. Getting inventorship right from the start is not just a filing formality: errors can invalidate an otherwise valid patent.

Contact Goldstein Patent Law to schedule a consultation. We’ll help you establish the right inventorship record, structure your application to hold up under examination, and build an IP strategy that protects your AI-assisted innovations for the long term.