It replaces jobs previously done by humans and creates ideas and products that would normally count as an invention, and yet, artificial intelligence (AI) cannot be an inventor. This was the decision of the United States Trademark and Patent Office (USPTO), which held that the term “inventors” in the United States Code only pertains to “natural persons.” This decision is in line with decisions out of the United Kingdom and European Union. Such a stance, while adhering to the plain statutory text, is seemingly at odds with the circumstances of the 21st century and has the potential to have grave effects on technological development and protection of intellectual property.
The creation and use of artificial intelligence has skyrocketed in recent years. In 2019, the AI software market grew by approximately 154% over 2018. There now exist driverless cars, AI has taught itself chess and reports continue to indicate that AI will disrupt labour markets well into the future. In essence, AI has become a mainstay in society and its influence and reach will continue to expand with every year. Yet, the USPTO decision and many others are not in line with the changing nature of technology and society.
Under the current consensus, how can it be possible for the ideas created by AI to be protected under a patent or a copyright? Some, such as Lance Eliot, have argued that the people who programmed the AI could just name themselves the inventor. However, this notion is flawed. If the entire premise rests upon the notion, as 35 U.S.C. § 100(f) states, that an inventor is someone who “invented or discovered the subject matter of the invention”, then can one really argue that the creator of the AI, who never conceived, in their mind, of the product the AI created, is really an inventor? It would seem not. The existing regime would open up so many avenues to the theft of ideas, that if directly created by a human being, would be legally protected through a patent or a copyright. Under this system, creators are left gravely unprotected.
The public would also suffer from the continued application of this belief that AIs cannot be inventors. A study found that weaker IP protections were correlated to lower investment levels in the manufacturing sector, a sector that has been shrinking in both Canada and the United States. A continued adherence to the status quo where products developed by AI would not be protected by a patent would surely deter investment and innovation, as those who would have invested had there been a patent would realize the lack of protection may lead to increased competition as it would deprive the owner of the AI, as s.42 of the Patent Act states, “the exclusive right, privilege, and liberty of making… the invention.” Hence, firms and individuals would realize the limited protection and lower possible payoffs as they would not have exclusivity in the creation of the product, which limit incentives for investment and innovation. And as the OECD noted, innovation is necessary for tackling many of the problems society faces today. Thus, the public may very well suffer from the possible decreased innovation and investment.
The movement towards recognizing AIs as inventors for the purposes of obtaining a patent must be centred on the change of the law through elected legislatures, and not by deviation from the statutory text by unelected members of agencies or the courts. This would entail a complicated and arduous process, which would surely bring about extensive lobbying, as did the issuing of the Special 301 Report in the United States. However, these effects on government business are surely worth it to bring about needed change in the IP regime. To go the latter route, that is, through statutory interpretation by unelected members of the USPTO with decisions coloured by a view to the current times, would be to risk the supplementation of the laws with personal interpretation of the law. In the words of United States Chief Justice John Roberts in King v Burwell, “In a democracy, the power to make the law rests with those chosen by the people.”
It is time to acknowledge that in the age of growing technological innovation and increasingly sophisticated AI, that AI can be an inventor for the purposes of obtaining a patent. The decision of the USPTO and others holding otherwise is at odds with the current breadth of technology in our society. Moreover, under this system, the ability for creators to have the ideas created by AI be properly protected is jeopardized and the potential for technological innovation that would benefit all realms of society is possibly diminished. Despite the time and effort that a change to the laws would require by elected governments, such a change is both necessary and worthwhile.
Nicholas Heinrich is a fourth year undergraduate student at the University of Toronto studying Political Science and Economics.