How AI Tools Impact Prior Art and Examination

How AI Tools Impact Prior Art and Examination

View Session Speakers

This summer the USPTO has been soliciting formal comments from the industry relating to the impact of artificial intelligence (AI) on prior art and patentability issues.  Among other things, commenters have made various suggestions on how the USPTO could protect against the use of unreliable AI-generated references where a human wasn’t involved in publishing or verifying the accuracy of the information.

All current categories of Section 102 require some type of human action to place the information in the public domain – e.g., filing a patent, publishing an article, placing an object in public use or on sale, or otherwise making it available to the public. Thus, a question arises about whether creation of AI-generated content will (or should) require sufficient human action to place the information in the public domain and qualify as prior art under Section 102?

Because AI systems are known to hallucinate, AI-generated content may or may not be credible. So, even if a “human action” threshold is agreed upon for AI-generated information to become prior art, will (or should) AI-generated content need to have a sufficient indicia of reliability before examiners give it consideration in evaluating questions of patentability?

These and other questions about what is and should be considered prior art in the age of AI will be discussed by the panel.

Survey

https://www.surveymonkey.com/r/How-AI-Tools-Impact-Prior-Art-and-Examination