07.07.2022

The Paper “Obvious Patents are not Obvious: AI for Detecting Weak Patents” submitted by Egbert Amoncio for the DRUID 2022 Conference (13.-15.06.2022, Copenhagen) was one of the three nominees for the DRUID 2022 Steven Klepper Award for Best Young Scholar Paper.

Paper: “Obvious Patents are not Obvious: AI for Detecting Weak Patents”

Abstract:  The U.S. Patent and Trademark Office (USPTO) should only grant protection to intellectual properties (IPs) that fulfill the non-obviousness requirement. However, despite being crucial in filtering out patents with dubious validity (i.e., weak patents), little academic research is devoted to detecting the presence of patents with questionable non-obviousness (i.e., obvious patents). I address this gap by proposing a framework to detect obvious patents by employing AI. I first validate that indeed AI- based indicator captured obvious patents through two face validity tests. Using this validated measure, I analyze the impact of a policy strengthening the enforcement of non-obviousness requirements (i.e., KSR) on patent examination process efficiency. Given that backlogs and long pendency have been a long-standing issue in patent systems around the world, especially in USPTO, investigating this previously overlooked policy impact could provide insights into solving such issues. Results show that the introduction of KSR causally increased the efficiency of examiners, enabling them to process obvious patents almost 2 months faster. Lastly, in probing heterogeneous policy impacts, results reveal that the ex-ante stringent and ex-ante thorough examiners are responsible for the increased efficiency. I outline policy implications and recommendations.

Information on DRUID conference can be find here.