In deciding precedents including patent applications related to machine learning and artificial intelligence processes, the Federal Circuit Lyromive Analytics, Inc. v. FoxCorp. , No. 2023-2437 (Fed. Cir. April 18, 2025) determined that applying a general machine learning process to a new data environment would be insufficient to grant patent eligibility without technical innovations in the underlying machine learning process. on the other hand recently The decision represents the Federal Circuit's first foray into the eligibility of the subject of machine learning innovation. recently The decision is consistent with the existing §101 jurisprudence of the Federal Circuit for software and computer-implemented inventions.
recently decision
Slumative Analytics, Inc. filed a lawsuit against Fox Corp. in the Delaware area, claiming patent infringement for four patents related to the use of machine learning to generate network maps and optimize scheduling of television broadcasts and live events. Fox acted to dismiss the claim because the patent was not able to state the claim on the grounds that it was ineligible under §101. Recent patent claims recited the unique application of machine learning processes to the generation of network maps and scheduling live events and scheduling broadcast programming, which in turn a claim for a patent was indeed eligible for a patent. The district court granted and appealed Fox's motion, confirming it by the Federal Circuit.
in recently Decisions, the Federal Circuit ruled that “arguing that it does not merely apply established methods of machine learning to new data environments” is not eligible for patents. recentlyslip op. In reaching that decision at 10, the Federal Circuit emphasized:
- The claimed inventions relied on general and traditional machine learning techniques and common computers and processors.
- Iterative training or dynamic adjustment is not a technical improvement, as it falls within the nature of machine learning.
- Claims cannot be disclosed how Machine learning technology achieves improvements.
- Restricting machine learning to specific uses or technical environment areas, or applying existing technologies to new databases will not be subject to patents.
- Use machine learning technology to speed and efficiency of processes Without improving The underlying machine learning process is do not have Awarded patent eligibility.
Takeout from recently decision
recently The decision emphasizes that even in new technology environments, the addition of known or common machine learning or artificial intelligence processes is insufficient to award eligibility under 35 USC §101. After surviving the challenge of §101 recentlyApplicants should consider and actively argue with, not merely increase in computational speed or efficiency directed from machine learning processes or new use cases for known machine learning processes.
Furthermore, the Federal Circuit has shown that “improvement to machine learning models” can grant patent eligibility under §101, but questions remain about its qualification as an improvement in the context of the machine learning process. The boundaries and boundaries of these improvements are defined as digestion and interpretation of the US Patent and Trademark Office, district courts, and the Federal Circuit. recently decision.
