The U.S. Patent and Trademark Office (USPTO) last week rolled out its first artificial intelligence (AI)-based project, one that aims ultimately to speed patent examination processes, said Tom Beach, Chief Data Strategist and Portfolio Manager at USPTO.
Speaking at the Veritas Public Sector Vision Day event today, Beach said the project took about nine months to develop, and tackles what he called a “really compelling case” for use of AI tech.
The USPTO project is employing AI tech to extract technical data from patent applications, and to use that to inform Cooperative Patent Classification (CPC) data that is reviewed by USPTO patent examiners as they evaluate patent applications received by the agency. If patent examiners find the AI-assisted CPC data to be valuable, they will end up flagging less CPCs for further investigation, and speed the overall evaluation process.
“That’s the ROI [return on investment] for this project,” Beach said, who indicated that the success of the AI project also may allow USPTO to reduce what it spends on third parties who are now involved in the process.
USPTO will spend the next year operationalizing the project and seeking to validate its results, Beach said. If those phases produce positive results, he estimated that full implementation of the project will take place after that.
The project, he said, is capable of handling the 2,500 patent applications per day that USPTO receives on average.