A Diverse Patent Portfolio Better Protects Artificial Intelligence Inventions

Executive Overview

IBM sued Zillow in September 2019, alleging infringement of seven IBM patents directed to artificial intelligence (AI) algorithms for estimating property value. 2:20cv851 (W.D. Wash.) The lawsuit targeted Zillow's Zestimate service, which estimates a house's value using downloaded images of the house and neighborhood. IBM brought the litigation after three years of licensing negotiations broke down between the parties. While the infringement case is still ongoing, the Federal Circuit issued a decision to an intermediate appeal seeking invalidity of several IBM patents. IBM v. Zillow Grp., Inc., 50 F.4th 1371 (Fed. Cir. 2022).

This litigation reveals several takeaways about how to build a robust patent portfolio to protect AI inventions. First, notwithstanding a speedy allowance and persevering invalidity challenges, proving patent infringement on a specific AI algorithm can be onerous, so including additional patent claims targeting ancillary features of an applied AI system can be advantageous. Second, AI patent claims benefit from unique claim drafting techniques, specifically by separating the claimed method steps of training the AI model and those executing the AI model. Third, because the AI industry has a tendency to openly share new ideas, promptly protecting a new AI algorithm in a patent application is important to protect against waiver of patent rights.