Historically, patent inventorship interpre tation focused on whether legal rights belong to individual inventors, orga nizations or other entities. But as artificial intelligence evolved, the debate shifted to whether inventions generated with AI involve ment qualify for patent protection and who should be recognized as the inventor.
The United States Patent and Trademark Office has specific requirements for patent ability and provides guidance on what qualifies as an invention. Patent protection is generally available when an invention is useful, adequately described, novel and non-obvious. Patentability excludes laws of nature, physical phenomena, abstract ideas and mere suggestions.
The USPTO and United States Court of Appeals for the Federal Circuit provide guidance on inventorship. Corporations, universities and sovereign entities cannot be named as patent inventors, although they may be listed as applicants or owners. Instead, an inventor must be a natural person who signif icantly contributes to the conception of the invention—that is, to the formation of the inventive idea itself.
People who collaborate on an invention may be co-inventors.
They do not need to work at the same time, place or contribute equally, but each must contribute to at least one patent claim.
A new inventorship issue has arisen regard ing contributions by AI to a conception of an invention. In the landmark case Thaler v. Vidal (2022), AI was the sole contributor to the conception of the invention and was listed as the sole inventor. The federal circuit decided that AI cannot be listed as an inventor on a patent—but did not fully address how inven torship should be determined when AI plays a substantial role in developing an invention.
So, what happens when a person uses AI to significantly contribute to the conception of the invention?
AI’s disruptive role
AI’s contributions to the conception of inven tions complicate traditional inventorship definitions.
When used as a supplemental tool, AI can streamline brainstorming and problem identi fication, helping human inventors identify areas of improvement they might not have consid ered before. At the other end of the spectrum, certain AI systems can autonomously gener ate solutions with limited human direction, using human prompting primarily to initiate the process.
The first scenario is relatively uncontro versial, because using AI as a research tool is generally accepted in modern patenting. The second scenario, however, raises concerns about whether a human or AI satisfies the conception requirement.
Under current understanding, if a human satisfies conception, then the human is the inventor and the invention is patentable. But what if AI conceived the invention?
Under Thaler, AI cannot claim inventorship. So, does the invention become unpatentable?
Several unresolved questions need to be addressed—such as the type and percentage of human contribution, the qualification of a mental act in AI conception, and the USPTO patenting AI-conceived inventions originated or patented in other countries.
Determining an allowable amount of AI contribution will likely stem from a third scenario, which lies between refining an already conceived idea and relying entirely on AI to invent on its own. The use of AI does not prevent patentability; however, the permissible level of AI contribution remains unsettled.
In 2025, administrative rulings by the Patent Trial and Appeal Board clarified that AI may be used as a tool during the inven tive process. However, the human inventor must make meaningful contributions to the invention and generally cannot rely solely on AI-generated output.
Current laws and guidelines do not empha size bright-line rules. However, the USPTO and courts are attempting to establish boundaries.
Other emerging issues
The evolution of AI use has spread worldwide. While Thaler pushed for AI inventorship in the United States, the patent applicant also submitted patent applications in other jurisdic tions—namely, the United Kingdom, Europe, Australia and South Africa.
The U.K. and Europe paralleled the United States and denied the applications. Australia and South Africa granted patent inventorship rights to the AI system, finding no limitation in their patent law language.
These differing approaches may create challenges for international patent strategies and could encourage applicants to consider jurisdictions with different rules regarding AI-related inventorship.
Outside the realm of patents, the U.S. Copyright Office repeatedly denied a copyright registration for a piece of visual art generated autonomously by an AI system (the “Creativity Machine”) because it lacked a human creator.
Thaler again took this issue to court and lost. The court’s decision in Thaler v. Perlmutter aligned with the findings and recommendations of the U.S. Copyright Office’s January 2025 report. The guidance provides that if AI generates a work entirely on its own, it enters the public domain and cannot be copyrighted.
However, the Copyright Office has indi cated that if a human provides significant creative input (e.g., heavily editing, arranging or combining AI elements with human creation), the human-created portions may be eligible for copyright protection.
These guidelines provide insight into the limits of AI-generated use. However, ambiguity remains regarding the degree of human involve ment necessary to obtain protection.
Who is the inventor?
The United States is at a legal crossroads with the use of AI to contribute to inventions.
One possible approach would allow broader patent protection for inventions developed largely through AI-assisted processes, poten tially expanding the range of innovations eligible for protection. This will support America’s position as one of the largest contrib utors of patents worldwide. Additionally, it broadens acceptance of AI-assisted patents that might otherwise be denied under Thaler.
Another approach would continue requiring substantial human involvement before patent protection is available. This path will follow the current trajectory of AI limitations given by the USPTO and U.S. case law.
Requiring significant human contribution and conception would maintain the core values of creativity and ingenuity that contribute to the understanding of an inventor. Further, requiring a human element will support ethical accountability of inventors, which AI systems consistently lack, and prevent the patent system from being flooded with AI-generated patent applications.
Regardless of the Unites States’ position, AI will continue pushing legal and intellectual boundaries. Before clear lines can be set, several unresolved questions need to be addressed— such as the type and percentage of human contribution, the qualification of a mental act in AI conception, and the USPTO patenting AI-conceived inventions originated or patented in other countries.
Until then, inventors, businesses and practi tioners should approach these legal gray areas with caution while courts and policymakers continue to define applicable rules.
Originally published by Inventors Digest.
The content of this article is intended to provide a general guide to the subject matter. Specialist advice should be sought about your specific circumstances.
