In response to recent advances in artificial intelligence (AI) technology, including automation and predictive analytics, the United States Patent and Trademark Office (USPTO) should reevaluate its patent examination policy from the perspective of economics, fairness, time, and transparency.
So proposes California Western’s Professor Tabrez Ebrahim in a recent scholarly paper he presented at the 2019 Georgia State University Law Review Symposium titled Automation & Predictive Analytics in Patent Prosecution: USPTO Implications & Policy. This paper advances the field of legal analytics specific to patent prosecution, which concerns the negotiation between an inventor and the U.S. government for the granting of a patent. The paper was accepted for publication by Georgia State Law University Review.
“In economic terms, artificial intelligence technology reduces the transaction costs of acquiring patents, and its use during the patent prosecution process affects the distribution of power,” says Professor Ebrahim. “The presence of constantly updated patent prosecution data and the generation of predictive models expands the information asymmetry between inventors and patent examiners.”
In his paper, Professor Ebrahim analyzes how the interactions between inventors and patent examiners will evolve in response to AI technology through an economic-based lens. “The private sector’s capabilities will soon outpace those of the USPTO,” says Professor Ebrahim. “In response, I suggest that the USPTO reevaluate patent policy and proactively consider developing a counteracting review board at the USPTO.”
Patent prosecution is an excellent case study for the use of AI technology. It is considered a mechanical and structured law practice where automation technologies can assist patent practitioners in drafting patent applications. “One particular software tool can now automate 90 percent of the patent application drafting process in certain technology areas, thereby reducing patent drafting from an average 25 hours to five,” says Professor Ebrahim.
Moreover, the vast volume of patent history data, which is constantly updated, can be analyzed to provide predictive analytics of patent examiners’ decision and allowance characteristics. “Data-driven analytics and predictive modeling can provide a competitive advantage to the inventor over the USPTO,” says Professor Ebrahim.
The paper is centered on economic perspectives and policy arguments. But, according to Professor Ebrahim, the practical ramifications of AI technology will affect the administrative efficiency of the USPTO and have a profound impact on the practice and the profession of patent prosecution. “Graduating law students, patent attorneys, and in-house patent counsel with an understanding of AI technology and skills in predictive analytics could distinguish themselves among their peers in the legal marketplace for jobs and for representing technology company clients,” he says.
Professor Ebrahim’s paper closes with patent policy considerations, which he introduces to California Western students in his doctrinal Patent Law class and his Patent Litigation & Strategy class.
For example, the courses include discussion of how other nations patent offices’ have responded to AI technology developments and guidance for students seeking internships and ultimately jobs in patent law practice.
“I introduce to students how the use of these new algorithmic processes can effect accountability, bias, explainability, and transparency,” says Professor Ebrahim. “These are topics that relate to policy levers and how the USPTO should operate. These policy considerations would impact the patent prosecution profession, the practice of patent law, as well as the interactions between patent attorneys and their clients.”
Read Professor Ebrahim’s entire paper, Automation & Predictive Analytics in Patent Prosecution: USPTO Implications & Policy here.