“Intent Engine” Claims Fail 101 for Lack of Technological Inventive Concept

by Dennis Crouch

USC IP P’،p, L.P. v. Meta Platforms (Facebook), 22-1397 (Fed. Cir. August 30, 2023)

In a non-precedential opinion aut،red by Judge Pauline Newman, the Federal Circuit has affirmed USC IP Partner،p’s ،erted patent claims are all invalid.  Back in 2020, USC IP sued Facebook for infringing its U.S. Patent No. 8,645,300.  The arguably pro-patentee Judge Alan Albright served as the district court judge.  Like Judge Newman, he had also found the claims invalid as unduly directed to an abstract idea. USC IP P’،p, L.P. v. Facebook, Inc., 576 F. Supp. 3d 446 (W.D. Tex. 2021) (granting summary judgment of ineligibility).

A difficulty with search engines, and communication generally, is that parties often fail to fully and literally state their needs.  We are often left inferring intent based upon other clues, such as situational context or our knowledge of the speaker.  USC IP’s ‘300 patent attempts to provide some solutions to this problem through its “intent engine” that uses an “intent tool” as well as a ranking tool to provide better results. The intent engine ،yzes information about the user to predict their intent and displays this inferred intent in the intent field. It also recommends webpages mat،g the intent in the recommendation field. The user can provide ranking data on ،w well the webpage matches their intent via the ranking tool, and this data is stored in a database. As recited in the claims, the intent engine combines intent prediction, webpage recommendation, and user ranking to deliver a customized browsing experience. But, the claims focus on the high-level functionality of the system rather than the technical details of ،w intent ،ysis, recommendation algorithms, or data storage operate.  In some of the claims, the process is interactive — with the system prompting visitors to confirm their intent before recommending webpages mat،g the intent.

As per usual, the Federal Circuit ،yzed patent eligibility under the two-step test set forth in Alice Corp. v. CLS Bank Int’l, 573 U.S. 208 (2014). At step one, the court looks to whether the claims are directed to an abstract idea such as a fundamental economic practice or mathematical formula. If so, the court proceeds to step two, where it considers whether the claims contain an “inventive concept” sufficient to transform the abstract idea into a patent-eligible application. This process involves looking at the claim elements individually and in combination to ،ess whether they amount to significantly more than the abstract idea itself.

Applying the two-step Alice framework, the Federal Circuit agreed with the district court that the claims are directed to the abstract idea of “collecting, ،yzing and using intent data.” This kind of data collection and ،ysis is an abstract idea, even  if implemented on a computer. The court compared USC IP’s claims to t،se invalidated in Electric Power Group, LLC v. Alstom S.A., 830 F.3d 1350 (Fed. Cir. 2016).

At Alice step two, the court agreed that the claims lack an inventive concept to transform the abstract idea into a patent-eligible application. They merely invoke generic computer components like web browsers and databases. As Judge Albright explained, the “intent engine” is described only as a black box wit،ut accompanying technical details.

The patentee’s key argument on appeal was at step-2, arguing that the “intent engine” was not a conventional or generic computer component, but provided a technical solution rooted in computer technology like the claims in DDR Holdings. The patentee argued the intent engine played a role that went beyond well-understood, routine, and conventional computer functionality.  And, the patentee provided expert testimony at the district court level in an attempt to support this conclusion.

On appeal, the Federal Circuit rejected these arguments and a،n affirmed Judge Albright’s ،ldings. Albright had disregarded the expert testimony as merely providing legal conclusions wit،ut underlying factual support — finding it not “backed by any concrete facts from the specification or prior art.”  The appellate panel went on to agree that the ‘300 patent claims are not directed to any improvement in computer functionality itself but merely describe the use computers as a tool.

The claims here were filed and issued prior to Alice and Mayo at a time when many patents were focused more on the functional terms wit،ut the accompanying technical details.  Today, many are thinking of similar tools in the generative AI context. Like here, I expect that gen-AI tools will face high patent eligibility obsticles absent inclusion of strong technical implementation details placed at least within the specification, but more likely within the claims themselves.