Artificial Intelligence and Access to the Patent System
55 Pages Posted: 29 Jun 2020 Last revised: 15 Jan 2021
Date Written: March 10, 2020
How likely is it that the average American will become an inventor? With a novel idea and hard work, it should be a possibility for all Americans. The data suggests otherwise. Most patents are obtained by inventors that work for large corporations. Small businesses, solo inventors, women, and minorities lag behind their counterparts in patenting. A common explanation for this phenomenon is that it is a “pipeline” issue. However, evidence suggests that the patent system is not accessible to underrepresented innovators for more problematic reasons.
At almost the same time as information about the concentration of patenting activity has garnered attention, the U.S. Patent and Trademark Office (“USPTO”) has become more vocal about its interest in Artificial Intelligence (“AI”). AI will transform how the USPTO examines patent applications. It may also transform how people invent. What has been absent from the conversation about AI and patenting is the negative effect AI has had when introduced in other social systems. For example, AI models can exhibit bias that concentrates power around incumbents. In addition, leaders overestimate the ability of AI models to solve human issues. This article refers to these problems collectively as “AI enthusiasm.”
This Article argues that AI enthusiasm threatens to make the patent system less accessible for underrepresented innovators. In response, this article presents a framework for improving access to the patent system given the emergence of AU. First, limits must be placed on AI-assisted examination informed by best practices that combat AI bias. Second, the USPTO should grant patents to inventions that are created with the assistance of AI only if the AI involved adheres to a set of best practices that reduce the chance of biased outcomes. Finally, true access involves removing obstacles to the innovation culture that has historically been closed to underrepresented inventors. Thus, AI tools should be deployed to assist underrepresented innovators in the patenting process. Collectively, these measures may provide U.S. innovators from all walks of life the opportunity to call themselves an inventor.
Keywords: Patent Law, Artificial Intelligence, Innovation, Bias, Inventors
Suggested Citation: Suggested Citation