Can Artificial Intelligence Be an Inventor on a Patent?

Under current U.S. patent law, artificial intelligence cannot be named as an inventor. The Patent Act defines an inventor as the “individual” who invented or discovered the claimed subject matter, and the Federal Circuit has held that an inventor must be a natural person—a human being.

That does not mean an invention developed with AI assistance is automatically excluded from patent protection. In November 2025, the USPTO rescinded its earlier AI-specific inventorship guidance and clarified that the same legal standard for determining inventorship applies whether or not AI was used. AI systems are treated as tools used by human inventors, not as inventors themselves.

For companies using AI in research, engineering, software development, product design, drug discovery, and other innovative work, that distinction creates the more important business question: Which humans actually contributed to conception of the invention being claimed?

Why Can’t AI Be Named as a Patent Inventor?

The Patent Act defines an “inventor” as the individual, or the individuals collectively in a joint invention, who invented or discovered the subject matter of the invention. The Act also generally requires each individual inventor or joint inventor to execute an inventor's oath or declaration in connection with a patent application.

The Federal Circuit directly addressed AI inventorship in Thaler v. Vidal. The patent applications at issue identified an AI system called DABUS as the inventor. The court held that the Patent Act requires inventors to be natural persons and therefore does not permit an AI system to be named as an inventor.

The USPTO’s current guidance follows that rule. It expressly states that only natural persons can properly be named as inventors and treats AI systems—including generative AI and other computational models—as tools that may assist human inventors.

Can an AI-Assisted Invention Still Be Patented?

Potentially, yes.

Using AI during the inventive process does not itself prevent patent protection. The USPTO’s current position is that ordinary inventorship law applies regardless of whether AI was among the tools used to develop the invention. There is no separate or heightened inventorship standard for AI-assisted inventions.

This distinction matters because sophisticated tools have long played roles in innovation. Researchers use laboratory equipment, engineers use simulation and modeling systems, and developers use computational tools. AI may be unusually capable of generating proposed solutions, but the inventorship inquiry still focuses on the human contribution to conception.

The better question for a company is therefore not merely whether AI was involved. It is who qualifies as a human inventor of the subject matter the patent application actually claims.

Why Does Conception Matter to AI Inventorship?

Patent inventorship is closely tied to conception. The USPTO’s revised guidance emphasizes that

invention under the Patent Act centers on the inventor’s conception and that the inventorship analysis for AI-assisted inventions should continue to focus on human contribution to conception.

That means inventorship is not simply a list of everyone who worked on a project. Someone does not necessarily become an inventor merely by supervising a team, funding development, testing an already-conceived idea, collecting data, or carrying out instructions.

AI does not change that basic framework. The company needs to determine what the human contributors conceived in relation to the subject matter ultimately claimed.

This also means inventorship can be claim-specific. As claims are drafted, amended, added, or removed during patent prosecution, the proper inventorship analysis may need to be revisited.

Does Writing an AI Prompt Make Someone an Inventor?

Not automatically.

A person who enters a prompt into an AI system does not necessarily become an inventor of everything the system produces. The relevant question remains whether the person’s contribution qualifies under ordinary inventorship principles.

A detailed prompt may reflect human conception of important aspects of a solution. A broad instruction asking an AI system to solve a problem may present a different situation. The legal significance depends on the human contribution to the claimed invention rather than on the mere fact that prompting occurred.

This is also an area where relying on outdated descriptions of USPTO policy could cause problems. The USPTO rescinded its February 2024 AI-specific guidance in November 2025 and now expressly applies the same longstanding inventorship standard to AI-assisted and other inventions. There is therefore no simple rule that “the person who wrote the prompt is the inventor.”

What If AI Produces an Unexpected Solution?

AI-assisted development can create harder questions when the system produces a solution the human user did not anticipate.

Suppose a researcher identifies a technical problem and asks an AI system to propose possible solutions. The AI generates an approach the researcher had not previously considered. The researcher recognizes its significance, modifies the proposal, tests it, and develops it into a commercially valuable technology.

Inventorship should not be resolved simply by naming the researcher because a human inventor is required. Nor should the company automatically conclude that the invention cannot be patented merely because AI played an important role.

Instead, the patent claims and the actual human contributions need to be evaluated. Depending on what is ultimately claimed, human conception before, during, or after the AI interaction may determine who qualifies as an inventor. That makes documentation increasingly important.

What If AI Generated the Entire Inventive Concept?

This presents the harder boundary.

If no natural person qualifies as an inventor of the claimed subject matter, current U.S. law does not permit the company to solve the problem by naming the AI system. Thaler holds that an inventor must be a human being, and the USPTO continues to apply that requirement to AI-assisted inventions.

Nor does owning or controlling the AI system necessarily solve the inventorship problem. Inventorship and ownership are different concepts. A company may acquire ownership of patent rights through assignments or other arrangements, but the application must still correctly identify the qualifying human inventor or inventors.

For businesses increasingly using autonomous AI systems in R&D, this distinction can become commercially important. A valuable technical output does not automatically become a patentable corporate asset merely because the company paid for the AI system or owns the resulting data.

Can a Company Be the Inventor?

No. A company can own patent rights, but it cannot itself qualify as the natural person who invented the claimed subject matter. In ordinary corporate innovation, individuals may qualify as inventors while assignments transfer the resulting patent rights to the business. AI-assisted development does not eliminate that distinction.

Companies should therefore keep two questions separate: Who invented the claimed subject matter? Who owns the resulting patent rights? Both can matter when filing applications, licensing technology, conducting diligence, financing a company, completing an acquisition, or enforcing patent rights.

How Does AI Affect Joint Inventorship?

AI-assisted projects may involve multiple people contributing at different stages. One person might identify a technical problem, another develop the AI workflow, another recognize the significance of an AI-generated result, and another make modifications that become part of the claimed invention.

Not everyone involved is necessarily an inventor.

Federal patent law recognizes that joint inventors do not need to work together at the same time, make the same type or amount of contribution, or contribute to every claim. Even so, each person named as an inventor must qualify under the applicable inventorship principles.

Companies should therefore avoid both extremes: naming everyone who participated in an AI-assisted project or automatically treating the person operating the AI system as the inventor.

Does AI Use Have to Be Disclosed to the USPTO?

The USPTO’s current inventorship guidance does not create a separate general inventorship standard simply because AI was used. The agency’s current AI resources state that the same legal standard applies to all inventions regardless of whether AI participated in the inventive process.

That does not make internal documentation unnecessary. A company may need development records to determine inventorship accurately, prepare a patent application, resolve contributor questions, support due diligence, or respond to a later dispute.

The practical objective is not to preserve every interaction with an AI tool merely because AI was involved. It is to retain enough information to understand the human inventive contributions when those contributions matter.

How Should Companies Document AI-Assisted Inventions?

Companies using AI materially in research and development should consider whether their existing invention-disclosure process captures enough information to evaluate human inventorship.

For potentially valuable inventions, useful information may include the problem the team was trying to solve, important human ideas developed before AI was used, significant prompts or instructions where relevant, important AI-generated outputs, human selection or rejection of proposed solutions, later modifications, testing, and the relationship between those contributions and the invention ultimately pursued.

The objective should not be to create a burdensome record of every prompt and response. Instead, companies should preserve enough information to reconstruct the development of important inventions so patent counsel can compare the human contributions with the claims being pursued.

Should Companies Update Their Invention Disclosure Process?

For companies using AI extensively, a modest update may be worthwhile.

Traditional invention disclosures already ask about contributors, the technical problem, the proposed solution, development history, and relevant dates. A company can supplement that process by asking whether AI played a material role and, when it did, what the human contributors did before and after the important AI output.

This does not create a special legal test. The purpose is practical: better information gives patent counsel a better basis for applying the ordinary inventorship rules accurately.

Why Does Correct Inventorship Matter?

Inventorship is more than an administrative label. Patent applications must identify the inventor or joint inventors, and the Patent Act generally requires individual inventors to satisfy applicable oath or declaration requirements.

Inventorship can also matter in ownership analysis, assignments, licensing, transactions, due diligence, disputes, and enforcement. AI therefore creates an internal governance issue in addition to an interesting legal question.

Companies using AI extensively for innovation should know who is responsible for identifying potentially patentable inventions, preserving meaningful development information, and involving patent counsel before the inventive history becomes difficult to reconstruct.

Could U.S. Law Eventually Permit AI Inventors?

Congress could amend the Patent Act, and future legal developments could change the governing framework.

As of August 2026, the Patent Act continues to define an inventor as an individual, Thaler v. Vidal holds that inventors must be natural persons, and the USPTO’s current guidance treats AI systems as tools rather than inventors.

Businesses should therefore build current patent processes around human inventorship rather than assuming future law will recognize AI itself as an inventor.

Practical Considerations

The practical lesson is not that companies should avoid using AI to innovate. AI-assisted inventions are not categorically excluded from patent protection, and the USPTO applies ordinary inventorship law to inventions developed with AI assistance.

Companies making substantial use of AI should instead make proportionate adjustments to invention management. Development teams should preserve enough information to identify meaningful human contributions, invention disclosures should flag material AI involvement when it affects understanding of the inventive process, and patent counsel should evaluate inventorship against the claims actually being pursued.

The goal is not additional bureaucracy. It is to avoid discovering after a technology becomes valuable that the company cannot reliably determine who conceived the claimed invention.

Key Takeaways

AI cannot currently be named as a U.S. patent inventor. Federal patent law and Federal Circuit precedent limit inventorship to natural persons.
Using AI does not automatically prevent patent protection. The same inventorship standard applies regardless of whether AI assisted in developing the invention.
Prompting an AI system does not automatically establish inventorship. The analysis focuses on human contribution to conception of the claimed subject matter.
Inventorship and ownership are different. A company may own patent rights, but only qualifying human inventors can be named as inventors under current U.S. law.
AI-intensive businesses should document important human contributions without creating unnecessary administrative burden. The objective is enough information to determine inventorship reliably when a potentially valuable invention arises.

Conclusion

Artificial intelligence can be a powerful tool for invention, but under current U.S. patent law it cannot itself be named as the inventor. The more useful question for businesses is whether the claimed invention resulted from human contributions sufficient to identify one or more natural persons as the proper inventors.

As AI becomes more deeply integrated into research and development, inventorship will increasingly become an invention-management issue as well as a legal one. Companies do not need an entirely separate patent system for AI-assisted innovation, but they may need better documentation and communication among technical teams, management, and patent counsel.

The companies best positioned to protect valuable AI-assisted innovations will be those that can determine what their AI tools contributed, what their people conceived, and how those human contributions relate to the patent claims they ultimately seek.

Additional Resources:

Related Industry : Technology, Software & AI

Related Practice Area : Patent Law

About the Firm

Klemchuk is a litigation-led, full-service intellectual property law firm serving sophisticated businesses, entrepreneurs, and investors. We help clients protect, commercialize, and enforce valuable intellectual property by combining deep IP experience, practical business judgment, senior-level relationships, and trial-ready litigation capability.

This article is provided for informational purposes only and does not constitute legal advice. Artificial intelligence tools may have been used to assist in researching, drafting, editing, or reviewing this content. The content is subject to human review, but AI-generated or AI-assisted content may contain errors or omissions. The appropriate legal strategies depend on the facts and applicable law. The law evolves, and this article likely will not be edited to reflect every change in the law. Laws may also differ or conflict between jurisdictions. This article may contain mistakes. For all these reasons, you should consult a competent attorney for legal advice and should not rely on this article as a substitute for advice concerning your particular circumstances.

In Depth Guide, Patents, Software & AI, IP PortfoliosKlemchuk PLLCSeptember 15, 2026Comment