USPTO Patent Classification and Search Systems Using AI


There is no doubt that artificial intelligence (“AI”) technologies have emerged in all aspects of society, such as online shopping, music streaming and social media. The United States Patent and Trademark Office (“USPTO”) even reported that patents incorporating AI rose from less than 5% in 1980 to more than 20% in 2018. Among the organizations that use AI are finds the USPTO itself. The USPTO has recently incorporated AI technology into patent examination, particularly with prior art search and patent classification. These changes can bring various benefits to patent applicants, although their implementation is behind the scenes of the patent examination process.

The considerable growth in the state of the art and the exponential rate of developing technologies have made prior art searches more difficult and burdensome for patent examiners. However, this research is essential for the quality examination of patent applications, so that patents are granted with a reliable validity that will last in any future litigation. To reduce the amount of resources spent on prior art research, the USPTO implemented an AI-based prototype research system. This AI system aims to help identify relevant documents and provide suggestions to reviewers for additional areas of research. This AI system is also configured to continuously internalize reviewer feedback, which will improve the technology over time. This prototype was presented to a select group of reviewers in March 2020. Since then, positive feedback has been received and the USPTO plans to incorporate this AI technology into its next-generation search tools for reviewers.

The USPTO has also developed an automatic classification AI tool that can classify patent applications according to the Cooperative Patent Classification System (“CPC”). This AI system identifies the claimed object, suggests CPC codes, and relates a specific object in apps to each specific CPC code. The USPTO implemented this automatic classification AI program in December 2020. As a result, the USPTO has seen a reduction in spending for CPC data acquisition. For reference, in 2015, the USPTO spent $ 95 million on a 5-year patent classification contract at the Serco Company.

The AI ​​technology that the USPTO intends to implement on a larger scale is also aimed at streamlining the patent application process, potentially making it faster and more cost effective for applicants. Especially with automatic classification AI technology, patent applicants may see a decrease in USPTO filing fees as the USPTO would no longer need expensive contracts for classification purposes. Alternatively, the USPTO could use the saved expenses and hire more examiners, which would reduce an applicant’s wait time for the exam. While a contracted classifier can take months or even over a year to classify patent applications throughout their submission queue, an automatic classification AI system will have the ability to classify patent applications. requests very quickly, even instantly upon submission. The USPTO could take advantage of this, for example, by reducing the time before a patent application is published. With AI technology for prior art search, claimants will benefit from a greater likelihood of obtaining the closest prior art to their invention at the start of a lawsuit and not after multiple administrative actions. Thus, with fewer office actions and relevant prior art being more likely to be identified at the start of patent examination, applicants can save time and money.

Overall, these AI technologies recently implemented at the USPTO increase the likelihood that patent applicants will obtain high-quality patents faster and at a more reasonable cost. Going forward, the USPTO plans to implement even more AI-based technologies, including an image search program for further prior art research. This could be particularly useful in the area of ​​design patent applications.

Co-Written by Amanda Metell, Summer 2021 Associate

© 1994-2021 Mintz, Levin, Cohn, Ferris, Glovsky and Popeo, PC All rights reserved.Revue nationale de droit, volume XI, number 189


Previous On campus and online: Harper is ready to serve you: Harper College
Next Roundup: Australia Seeks Developer of COVID-19 Vaccine Certification Applications; South Korean blockchain company releases source code for COVID-19 digital pass and other briefs