Further to Woods Rogers’s recent e-alert, the U.S. Patent & Trademark Office (USPTO) issued guidance on the patentability of inventions developed using artificial intelligence (AI). The guidance—which has sparked a flurry of reactions from global companies and trade organizations, as summarized below—states that patents are available for inventions created with AI assistance, provided a human has made a “significant contribution” to every element. The USPTO emphasizes that the “inventorship analysis should focus on human contributions” as patents function to incentivize and reward human ingenuity.

Questions remain on the inventorship boundaries when using an AI engine to supplement or complement the innovation process. The USPTO guidance expressly notes that “maintaining ‘intellectual domain’ over an AI system” does not make a person an inventor. Similarly, merely recognizing a problem and presenting the problem to an AI system does not constitute a significant contribution to support a claim of human inventorship. On the other hand, significant contribution can be shown by formulating AI engine prompt(s) to “elicit a particular solution from the AI system.”

The guidance has been met with a mixed response. Some, like Google, have praised the USPTO for bringing certainty to the issue of AI in patents. The Intellectual Property Owners Association and the American Intellectual Property Law Association, two highly respected industry groups, have also broadly embraced the USPTO’s approach. 

Critics, on the flipside, argue the guidance could discourage AI adoption and put patents at risk. The Council for Innovation Promotion, a group founded by former USPTO directors and Federal Circuit judges, has called for the guidance to be rescinded or substantially revised.

The Biotechnology Innovation Organization expressed concerns about the United States becoming an outlier, where inventors might believe that using AI could hinder the ability to get a patent. Novartis echoed these sentiments, stating that the idea that a novel, non-obvious invention could be denied a patent because the inventor used AI is "both factually and legally wrong, and ultimately amounts to the wrong policy for America.” The Institute of Electrical and Electronics Engineers argued the USPTO should not be trying to determine if an inventor was a human or a machine.

The issue remains hotly debated in real-time, and additional guidance from the USPTO is forthcoming. Woods Rogers’s multidisciplinary AI team will keep you apprised of related developments and the business impact of using AI and machine learning.