USPTO Signals Increased Patent Protections for AI and Machine Learning

John previously wrote about an internal memo at the US Patent and Trademark Office (USPTO). That memo encouraged USPTO patent examiners to exercise greater care before deeming AI and machine learning patent-ineligible abstract ideas. Now, a recent reversal from the USPTO shows that reasoning in action.  In Ex parte Desjardins, Appeal 2024-000567, the patent examiner initially found that the applicant’s machine learning methodology was unpatentable. However, on appeal, a panel headed by the USPTO director reversed this decision. A recent Thompson Hine memo discusses what this means for the patent eligibility of AI and machine learning systems:

“The reversal of the new ground of rejection signals that Director Squires is interested in expanding the patent eligibility of artificial intelligence and machine learning claims, as well as a subtle but meaningful shift at the USPTO regarding the examination of these claims. This is encouraging news for innovators in artificial intelligence and machine learning technologies.”

This expansion of patent protection is in line with the administration’s focus on AI expansion. Intellectual property law is often viewed as a barrier to AI development. When the administration issued a call for public comments on laws that hinder AI development, I noted that IP law may be a prime target of those reforms.  We’re now seeing patent law expanded to offer more protections to AI and machine learning applications. We’ll see if the US Copyright Office follows suit and changes its position to be more AI-friendly as well.