Innovation at the USPTO! New Agentic AI and Image Search AI Features Seek to Improve the Trademark Application and Examination Process for Applicants and Examiners
The noise around AI adoption in the legal industry has reached deafening levels. Seemingly every other week one of the major players such as Anthropic, Harvey, or Legora announces a new model, plugin, platform, or tool to improve legal workflows, or a law firm announces firmwide adoption of the same. In some cases, law firms have even announced their in-house development of bespoke AI-tools. As it turns out, the United States Patent and Trademark Office (“USPTO”) has decided to join in on the fun: this past April the USPTO announced the launch of several new AI features to facilitate an easier and more efficient trademark search and application process.
The first one is a feature in the USPTO’s trademark search system that allows users to search for trademarks that are similar to their mark by uploading an image of their mark. The system then identifies trademarks that contain design elements that are similar to the uploaded image of the user’s mark. While this is a step in the right direction to modernizing the USPTO’s search platform, the announcement signals that this feature is not fully developed because the USPTO is telling users to try their “new beta features.” Perhaps hedging further on the tool’s capabilities, the announcement states that the feature should be viewed as a supplemental search strategy that “doesn’t replace a traditional clearance search.” In any event, that applicants and registrants can now conduct image searches the same way as they can on a search engine is a positive development, as it provides users with another way to cross-check for conflicting marks that does not involve using design search codes.
The second is a feature that addresses mark descriptions and color claims for design marks, which can be a highly technical and time-consuming aspect of applying for a trademark with design elements that can cause delay in the application process if it is not done correctly. Specifically, every significant visual element such as shape, color, and/or size must be described to claim sufficient protection over those elements. To alleviate the frustrations associated with this process, the USPTO’s new feature allows applicants for design marks to upload images of the mark. The AI tool then makes suggestions for the mark description and color claims. Upon close review for accuracy, applicants can then accept, reject, or modify the AI tool’s suggestions and, to the extent not rejected, import them into the trademark application. Clearly, the USPTO is leaning into the use of image searching and rightfully so. Trademarks often include design elements or logos that can be difficult to describe but are critical to defining and protecting trademark rights. This feature provides a straightforward and efficient method for applicants to capture the design elements of their mark in a less tedious manner that is more natural, efficient, and (ideally) accurate.
The third is an agentic tool (because why not join the agentic wave!) that can be used to immediately assign international classes to unclassified applications, and to add design search codes and pseudo marks to applications so that they are searchable for trademark examiners and attorneys alike. The USPTO has dubbed this tool the Trademark Classification Agentic Codification Tool, or “Class ACT.” In the announcement for Class ACT, the USPTO acknowledges that the process of adding design search codes, pseudo marks, and international classification to applications for marks with logos, designs, unconventional spelling, and/or no international classes designated bogs down examiners, who rely on these items to conduct searches in connection with examinations. The Class ACT also benefits the public in that it reduces the timing of classification and design search coding from five months to minutes, and allows the trademark examination process to proceed more efficiently.
While the USPTO’s AI improvements do not rank among the flashiest that have recently been announced in the legal tech world, they certainly demonstrate that the USPTO is embracing generative AI to improve speed and efficiency for trademark applicants. We expect this trend to continue, as the USPTO has stated that “[m]ore AI trademark solutions are coming soon.” One can only hope for a Harvey or Legora for trademarks next!
