U.S. patent laws are designed to promote science and the useful arts. They grant temporary monopoly rights to inventors in order to incentivize inventive activity. In the United States, patent rights revolve around the inventor. However, what happens when an Artificial Intelligence (AI) machine invents? Who deserves monopoly rights to the invention? Who will be incentivized by such monopolies? Do U.S. laws protect companies’ investments in AI? In 2019, for the first time in history, an AI machine called DABUS was listed as an inventor on two U.S. patent applications. The United States Patent and Trademark Office denied the applications, saying that inventors must be human. A federal district court affirmed the denial, as well as the Federal Circuit Court of Appeals, saying that the Patent Act “unambiguously” requires an inventor to be a natural person. Under the current statutory scheme, the two main problems AI-generated inventions present are inventorship and ownership. AI is already part of modern inventive processes, such as the development of vaccines or safety equipment. AI’s prominence will only increase because AI is able to do things humans simply cannot and thus will dramatically improve our lives. Therefore, the United States needs to ensure companies are properly incentivized to develop and use AI. Currently, U.S. patent law comes up short. If AI-generated inventions are unpatentable, what will incentivize companies to invest significant resources into inventive AI? Will countries such as South Africa, which allowed the DABUS inventions to be patented, put the United States at an inventive and technological disadvantage? To solve the problems of inventorship and ownership, I propose the creation of a sui generis category of invention whereby AI-generated inventions without a human inventor can receive patent protection and the company that employs the AI becomes the “inventor” and owner under a work-made-for-hire-type model.

First Page