The Federal Circuit recently held as a matter of statutory interpretation that an artificial intelligence system cannot be named as an inventor on a US patent application.
This holding, which effectively excludes AI systems from the category of “individuals” eligible to be named as inventors, may complicate the intellectual property strategies of innovators who use advanced AI for research and development. Here’s what happened and why it matters.
The Federal Circuit was asked to determine whether an AI system called DABUS could be named as the inventor on two separate patent applications. The first disclosed a light source that was calibrated with a specific frequency corresponding to, among other characteristics, certain human brainwave activity.
The second disclosed a design for a beverage container that, rather than being smooth like ordinary containers, had a complex surface structure based on fractal geometry.
The circumstances surrounding the creation of these two inventions was highly unusual. Steven Thaler, the named plaintiff, and creator of DABUS asserted that DABUS generated both of the inventions without any contribution from Thaler and, further, that any person having skill in the art could have taken DABUS’ output and reduced the ideas to practice.
Interestingly, Thaler’s assertion that DABUS independently “conceived” of these inventions, which is traditionally considered to be the mental part of the inventive act, was not disputed in the record.
“Conception” is ordinarily the touchstone of inventorship. Consequently, if DABUS had been a natural person, there would have been little dispute that he or she should be named as the inventor.
The Federal Circuit, however, found as a matter of statutory interpretation that an AI system is simply not eligible to be an inventor under the US patent statutes.
Patent Rights Originate With Inventor
The unequivocal holding that AI cannot be named as an inventor on a US patent application may become more important as increasingly sophisticated AI systems are used to generate novel and valuable inventions.
One of the foundations of patent law is that patent rights originate with the inventor. This well-known principle is reflected in the contracts that commonly require employees to assign inventions to their employers.
Thaler’s holding to preclude AI systems from being listed as patent inventors may create a category of orphan inventions. If an AI independently “conceives” of a patentable invention, just as DABUS purportedly did, no patent can issue with the AI named as the inventor.
At the same time, if a patent substitutes the name of a natural person for the AI as the inventor, when the person did not contribute to the conception of the invention, that patent would be subject to invalidation for naming the wrong inventor. This conundrum may leave inventions independently “conceived” by AI ineligible for patent protection.
Limited Impact for Now
For now, Thaler’s impact may be limited. The Federal Circuit stated that its decision does not address patent protection for inventions made with the assistance of AI, which is likely the far more common scenario at the present time.
If a researcher uses AI as a tool, the use can be analogized to using a computer to conduct complex calculations, data analysis, or simulations in which case the researcher directing or using the AI is likely to be the appropriate inventor.
As AI becomes more sophisticated, however, more research may resemble the facts considered in Thaler, in that the AI machine may actually “conceive” of potentially patentable inventions.
Congress or the Supreme Court could fill the Thaler-sized hole in inventorship eligibility by amending or interpreting the patent statutes to recognize that a natural person controlling, programming, or providing input to an AI is considered the “inventor” for the purposes of applying for patent protection on inventions potentially “conceived” by AI.
Unless and until that occurs, however, patentees may need to be cautious in their IP strategies regarding AI-generated subject matter.
For inventions that cannot readily be reverse engineered, AI-created advances may be subject to protection as trade secrets. For inventions that can be easily copied, such as DABUS’ fractal beverage container design, patent protection may remain the only viable form of IP protection.
In those circumstances, inventors and their patent attorneys will need to identify any AI-generated subject matter. They should document the ways in which that subject matter could be considered as being produced under the direction or input of the natural person to be named as inventor—which could render the AI nothing more than a research tool.
Finally, they need to work together to draft claims that cover ideas conceived by the human inventor.
Companies may also want to consider strategies based on a patchwork of patent and trade secret protections to encourage investment in groundbreaking inventions developed using AI that, under Thaler, cannot be protected by the traditional step of seeking patent protection.
Unless the Supreme Court or Congress step in to allow an AI to be designated as an inventor or the natural person controlling, programming, or providing input to the AI to be the inventor of any AI “conceived” invention, such strategies may become increasingly important as the capabilities of AI grow.
This article does not necessarily reflect the opinion of The Bureau of National Affairs, Inc., the publisher of Bloomberg Law and Bloomberg Tax, or its owners.
Robert A. McFarlane is a registered patent attorney and partner at Hanson Bridgett. He co-chairs the firm’s intellectual property practice, has been litigating patent cases in jurisdictions across the US for 25 years, and teaches patent law as an adjunct professor at the University of California, Hastings College of the Law.
Rosanna W. Gan is a senior counsel at Hanson Bridgett is an experienced patent litigator who focuses on patent and IP litigation and on complex appellate matters.