New U.S. Patent and Trademark Office (USPTO) Director John Squires was sworn into office on September 23. In his welcoming remarks, he stated that “every piece of IP we put into circulation is a potential job, a new business, a competitive advantage, or an investible asset … and yet another win for … society.” He characterized the role of the USPTO as “not to dispassionately say ‘no,’ but to engage and help find the ‘yes’ that is properly bounded by law.” Finally, he stated that the doors of the USPTO must be equally open and promised to ensure “that those doors are wide enough to accept whatever the on-rush new technology provides, wherever the puck is going.”
A day after his swearing in, Squires intimated at where he believes “the puck is going” by conducting apatent signing ceremony for two handpicked patents: one from distributed ledger/crypto technologies and another from medical diagnostics. By picking these two patents, Squires stated that he wanted to send a clear message: “the USPTO is open for business – especially for technologies of tomorrow.” Squires then listed other technologies, “from crypto and AI, to quantum computing and diagnostics,” that provide “breathtaking opportunities for invention and investment” and assured inventors and entrepreneurs everywhere: the USPTO is your partner.
Later that same week, Squires issued a rare appeals review panel (ARP) decision to address artificial intelligence.
The ARP Decision
The ARP decision involved patent claims related to training machine learning models. The ARP was convened to review a board decision entering a new ground of rejection on claims for lack of subject-matter eligibility. In the new ground of rejection, the board determined that the claims recited a mathematical concept (one of the three enumerated abstract ideas) and failed to include an additional element that integrated the abstract idea into a practical application.
The appellant disputed the board’s determination that the claims failed to include an additional element that integrated the abstract idea into a practical application, arguing that the “claimed subject matter provides technical improvements over conventional systems by addressing challenges in continual learning and model efficiency by reducing storage requirements and preserving task performance across sequential training.” The ARP agreed with the appellant.
Specifically, the ARP determined that at least one limitation of the claims reflected “an improvement to how the machine learning model itself operates, and not, for example, the identified mathematical calculation.” Based on this determination, the ARP determined that although the claims may recite an abstract idea, they are not directed to an abstract idea and are therefore patent eligible.
Addressing the board decision, the ARP highlighted the “overbroad reasoning of the original panel,” and acknowledged that this could be “understandable given the confusing nature of existing § 101 jurisprudence, but troubling, because this case highlights what is at stake.” The ARP asserted that the board’s analysis effectively “categorically exclude[ed] AI innovations from patent protection” which, in the ARP’s view, “jeopardizes America’s leadership in this critical emerging technology.” The ARP then stated that this case was useful in highlighting the appropriate tools for limiting patent protection to its proper scope, namely 35 U.S.C. §§ 102, 103, and 112. The panel noted that “these statutory provisions should be the focus of examination” – implying that Section 101 should not.
Takeaways
The ARP decision affirmatively pushes against using subject-matter eligibility as the primary tool for rejecting the patentability of an invention. First, the ARP decision recognizes that AI-related inventions are outside the abstract idea framework because “improvement[s] to how the machine learning model itself operates” are sufficient to overcome the subject-matter eligibility bar to patentability. Second, it explicitly reduces the focus of subject-matter eligibility during examination by calling for a shift back to the traditional, and appropriate, statutory provisions (Sections 102, 103, and 112) for limiting patent protection to its proper scope.
Taken in context with Squires’s preceding comments, the ARP decision appears to use AI as an example of how the USPTO’s treatment of emerging technologies, in general, may shift toward a more patentee-friendly approach moving forward.
Impact on Patent Strategies
The ARP decision signals a significant patentee-friendly shift in the patentability of emerging technologies and suggests a more aggressive patent filing strategy for emerging technologies, including artificial intelligence, medical diagnostics, distributed ledger, and crypto-related technologies.
The ARP decision also suggests that expanded coverage of emerging technologies focusing on improvements to the technology itself (such as improving the speed or accuracy of a machine learning model) could now be just as effective as the traditional approach of emphasizing improvements to computer technology (such as improving the speed or processing efficiency of a computer). With this in mind, new patent applications should be drafted with consideration of improvements to both computers, consistent with earlier strategies, and the emerging technology itself.
If you have any questions, or would like additional information, please contact one of the attorneys on our Intellectual Property team.
You can subscribe to future advisories and other Alston & Bird publications by completing our publications subscription form.