New U.S. Patent and Trademark Office (USPTO) Director John Squires has been on the job less than a month, but he has already indicated a significant shift in the USPTO’s willingness to extend patent protection to artificial intelligence (AI) related inventions.
On September 26, Director Squires, and the Appeals Review Panel (ARP), vacated the decision of the Patent Trial and Appeal Board (PTAB) panel to reject a Google AI-related patent application on §101 grounds. The ARP includes the Director, the Commissioner for Patents, and the Chief Judge of the PTAB.
Google filed a patent application in January 2019 addressing "catastrophic forgetting" in machine learning models, proposing methods to preserve knowledge of earlier tasks during training. The application was ultimately rejected on §103 obviousness grounds, and the PTAB panel reviewing the patent examiner’s decision later added a §101 rejection, deeming the claims abstract and lacking technological improvement under the famous two-part Alice patent-eligibility framework. The PTAB's panel decision was influenced by the Federal Circuit’s Recentive Analytics ruling, which held that applying established machine learning methods to new data environments without model improvements is not patent-eligible.
In recent years, a large number of AI-related patent applications have been rejected out of hand on §101 grounds. However, Director Squires indicated he takes a dim view of this approach.
In the September 26 ARP decision, Director Squires wrote, “Categorically excluding AI innovations from patent protection in the United States jeopardizes America’s leadership in this critical emerging technology. Yet, under the panel’s reasoning, many AI innovations are potentially unpatentable – even if they are adequately described and nonobvious – because the panel essentially equated any machine learning with an unpatentable ‘algorithm’ and the remaining additional elements as “generic computer components,” without adequate explanation.”
More specifically, the ARP supported its patent-eligibility decision with the Federal Circuit’s Enfish, LLC v. Microsoft Corp. case stating that the claims at issue recite “improvement to how the machine learning model itself operates[.]” He also commented that the §102 (novelty), §103 (obviousness), and §112 (written description, definiteness, and enablement) standards are the appropriate tools to limit patent protection to its proper scope.
Although the Federal Circuit Court of Appeals may weigh-in further on future AI-related cases, this ARP decision certainly is good news for companies innovating in the AI space in various industry sectors where AI is playing a more significant role. Moving forward, these companies appear more likely to gain patent protection for AI-related developments, especially where these developments are focused on improvements to the models or machines themselves.
Click here to read the Appeals Review Panel’s full decision in Ex Parte Desjardins et al.
[View source.]
