Though we recently touched on the U.S. Court of Appeals for the Federal Circuit ruling that only natural beings, which do not include artificial intelligence, can be an inventor under the U.S. Patents Act, the decision merits additional discussion.
Defining artificial intelligence (AI) can be a challenging task due to the expansiveness of its reach and evolving nature if its application. At its core, AI is the use of computer systems to perform tasks which historically have required human intelligence. Thus, AI can interpret and learn from data, creating alone, using what it has learned from this data, without additional human input. Such machine learning is further refined into deep learning, in which artificial neural networks—algorithms which copy how the human brain works—learn in a manner that mimics how humans do but by processing vast amounts of data. This neural networks approach of having computers “learn” as the human brain does differs from the more traditional rules-based computer programming, in which a computer is programmed with the explicit rules it will follow. Even as AI-powered computers begin to mimic human thinking, going beyond following predetermined tasks and routines, it’s not surprising that while AI is good at rules-based thinking, it cannot mimic certain functions of the human brain yet.
That “yet” is key—the legal status of whether AI can be an inventor is only beginning to be addressed in courts around the world, but the implications will be increasingly meaningful as the technology progresses and legal standing evolves. Consider the adjacent question of personhood: To date, unlike a corporation, AI does not have personhood standing, yet one robot, Sophia, has been granted citizenship by Saudi Arabia in October of 2017. Built by Hong Kong-based company Hanson Robotics, Sophia is designed to look and act like a human being, but only has human rights in Saudi Arabia. Further, AI in its current form is still performing based on rules created by humans.
Looking for a Beachhead
Stephen Thaler has filed a number of cases in various jurisdictions globally alleging that his DABUS AI system has created patentable intellectual property. Thaler has asserted that he creates and runs this AI system which can—alone—generate patentable inventions.
His most recent action was against the U.S. Patent and Trademark Office (PTO) and the court’s ruling came down resoundingly against him. (He has also lost similar actions in the European Union, the United Kingdom and Australia, winning only in South Africa.)
According to the court, “This case presents the question of who, or what, can be an inventor. Specifically, we are asked to decide if an artificial intelligence (AI) software system can be listed as the inventor on a patent application.”
The Court then turned to similar analysis undertaken by the PTO, who concluded that “the Patent Act defines ‘inventor’ as limited to natural persons; that is, human beings.” Thaler’s initial patent applications had not listed any human as an inventor; his DABUS system was listed as the sole inventor on his applications. The PTO determined that his applications lacked a valid inventor and were therefore incomplete.
As a result of this analysis, the PTO denied Thaler’s patent applications. Thaler had challenged this determination in court, appealing a lower court denial of his claim, hence the hearing in an appeals court.
According to the Court in this ruling, “the Patent Act expressly provides that inventors are ‘individuals.’” With the passage of the Leahy Smith America Invents Act in 2011, the Patent Act has defined an “inventor” as “the individual or, if a joint invention, the individuals collectively who invented or discovered the subject matter of the invention.”
Individuals, Oaths and Plain Meaning
Interestingly, there is no definition of the term “individual” in the Patent Act. However, according to the Court in this case, “as the Supreme Court has explained, when used ‘[a]s a noun, ‘individual’ ordinarily means a human being, a person.” [Mohamad v. Palestinian Auth., 566 U.S. 449, 454 (2012)]. And, the Court further added that when we use the word in common usage, or under its common definition, individual means a (natural) person. The Supreme Court has upheld this definition, unless Congress has clearly indicated otherwise—not present in this case. And, according to the Court here, “to the contrary, the rest of the Patent Act supports the conclusion that ‘individual’ in the Act refers to human beings.”
The PTO also requires individuals to submit an oath with their patent application that “such individual believes himself or herself to be the original inventor or an original joint inventor of a claimed invention in the application.” The Court did not believe that AI could provide such an oath given the uncertainty regarding AI’s ability to form a belief—a rationale that could preclude AI from being deemed an inventor unless full sentience is achieved. The Court noted that its ruling was further supported by its own prior precedent.
And, while the Court did note that statutes can be open to multiple reasonable readings, that principle was not applicable here as the statute in question was so clear in its meaning. Thus, the Court stated in its opinion that, “This is a case in which the question of statutory interpretation begins and ends with the plain meaning of the text.” Under the Patent Act, “individuals” means human beings. Further from the Court, “When a statute unambiguously and directly answers the question before us, our analysis does not stray beyond the plain text. Here, Congress has determined that only a natural person can be an inventor, so AI cannot be.”
The court also added that Congress had passed the Patent Act, and it was within its powers to pass a new law determining that others beyond those deemed a “natural person” could be inventors under the Patent Act. (Notably, the Court also assessed costs against Thaler.)
The case reads very cut and dry. According to the Court, this issue is one which must be legislatively addressed, as the text of the Patent Act seems quite clear, at least according to this Federal Court. However, while the issues in this case may have been amenable to resolution based on a relatively straight-forward analysis of the statutory language, future cases will undoubtedly be more complex. As an example, there are still open questions as to what extent a human can utilize AI tools during the inventive process and still qualify as an inventor. As AI becomes more sophisticated and can arguably iterate with little to no human intervention, these issues will become more pressing. In the meantime, at least in the eyes of the PTO and U.S. law, only the legislative branch would be able to give AI legal standing as an inventor. AI will continue needing human beings, at least for a while longer.
The Musical Twitter Bot: Who Has the Copyright for A.I.-Facilitated Works?
Deepfakes, Privacy Rights and an AI-Powered Blurring of the Lines