To date, there have been relatively few patent lawsuits relating to artificial intelligence. With the emergence of various artificial-intelligence-based software entering the market, it will be interesting to see how courts interpret artificial-intelligence-based patents and the issues that emerge from the court proceedings.  The following set of artificial-intelligence-related cases are from district courts of different geographic locations. 

Health Discovery Corp. v. Intel Corp.

577 F. Supp. 3d 570 (W. D. Tex. 2021)

In this case, Intel filed a motion to dismiss Health Discovery Corp (HDC)’s complaint on the grounds that the claims are invalid under 35 .U.S.C. § 101. HDC’s complaint alleges that Intel infringed HDC’s patents related to using learning machines, such as Support Vector Machines (SVM), to identify relevant patterns in datasets and to identify a selection of features within the datasets that best enable data classification. The asserted patents were directed to feature ranking, selection, and reduction using SVM to facilitate a Recursive Feature Elimination (RFE) process on a large dataset.

Under step one of the Alice test, the Court reasoned to follow the guidance provided from prior cases that analyzed patents with similar subject matter. After selecting two similar cases, the District Court asserted that the specification merely describes improving a mathematical analysis used by conventional systems. The specification explained how conventional systems reduce a feature size in data sets by ranking and eliminating features according to correlation coefficients. Similarly, the asserted patents involve ranking and eliminating features using SVM-RFE, which the court characterized as a purportedly novel but mathematical technique. As such, the District Court reasoned that the claims are directed to an abstract mathematical concept of SVM-RFE. 

Under Alice step two, the District Court stated that HDC’s complaint failed to sufficiently allege an inventive concept. The Court stated that improving data quality was an unpersuasive argument, and the Court also considered the fact that the claims were not limited to a particular field of invention. Accordingly, the Court granted the motion to dismiss under 35 U.S.C. § 101. 

Pavemetrics Sys. v. Tetra Tech 

2021 U.S. Dist. LEXIS 117651 (C.D. Cal. 2021)

In this case, Tetra Tech moved for a preliminary injunction to enjoin Pavemetrics from importing, using, and selling their “Laser Rail Inspection System” (LRAIL) products on the grounds of patent infringement. Tetra Tech’s ‘293 patent generally relates to a three-dimensional railway track inspection and assessment system that collects and processes data during and/or after a high speed assessment of a railway track. In 2018, Pavemetrics began developing AI-based deep learning algorithms using convolutional neural networks to identify defects in railway tracks in their LRAIL products.

Ultimately, the District Court concluded that Tetra Tech could not demonstrate a likelihood of success for a preliminary injunction because of substantial questions regarding infringement and invalidity. With regard to infringement, the Court noted that substantial questions remain because Tetra Tech relies on the “gradient neighborhood” limitation being reflected on a prior design of Pavemetrics’s LRAIL product from two years prior to the issuance of the asserted patent. At that time, Pavemetrics significantly changed how LRAIL processed data when it switched to detecting missing features using deep neural networks. The Court agreed that Tetra Tech had not provided sufficient evidence to indicate Pavemetrics’s “deep neural network” design within the current product meets the “moving gradient neighborhood like a sliding window over the 3D elevation data using the processor,” as recited in claim 1. 

With regard to invalidity, the Court determined there are substantial questions because Pavemetrics alleged that one of its prior designs anticipates claim 1 of Tetra Tech’s ‘293 Patent. For these reasons, the Court denied the motion for preliminary injunction.

IBM v. Zillow Grp., Inc.

2022 U.S. Dist. Lexis 41831 (W. D. Wash. 2022)

In this case, Zillow filed a motion to dismiss IBM’s claims of patent infringement on the grounds of ineligible subject matter under 35 U.S.C § 101. The ‘676 patent generally relates to a method of annotating response sets via an adaptive algorithm and the supplied annotations are used for a visualization system that present resource response results. In the specification, the ‘676 patent states that the “system discovers contexts and context attributes among users which can be used predictively,” by using “a highly specialized and optimized combination of supervised & unsupervised logic along with” automated entry of learned results. Col. 19, Lines 39-44.

Although the specification of the ‘676 Patent appears to describe improvements relating to computer and/or search engine functionality, the Court focused on the subject matter recited in the claims for step one of the Alice inquiry. The Court found that the claimed subject matter was directed to an abstract idea because the claim language was result-oriented and recited a process that could be performed with a pen and paper. The claims failed to recite the inner functionality of the invention. 

Under Alice step two, the Court concluded that the claims failed to satisfy the second Prong of the test for a couple of reasons. First, the specification failed to provide a description of the alleged inventive concepts offered by IBM. Second, the Court determined that the claims in the ‘676 patent do not provide a specific, discrete implementation of the abstract ideas for applying and ordering an annotation function, mapping the user context vector with the resource response set, or generating an annotated response set. Accordingly, the Court concluded the ‘676 Patent is invalid under 101 and the related patent infringement count is dismissed for failing to state a claim upon which relief can be granted.