Intellectual Property Report

Examining the Level of Detail Required to Plead a Claim Under the Doctrine of Equivalents
Katherine Corry
Patentees face many challenges in trying to prove literal infringement, particularly when dealing with competitors who make trivial modifications or superficial changes to the patented invention. In such situations, the doctrine of equivalents plays an important role in buttressing the scope of protection afforded by a patent and in preventing subversion of valuable rights through what amounts to nothing more than a misappropriation in disguise. At the pleading stage, patentees often attempt to avoid early dismissal by pleading infringement under the doctrine of equivalents, either alone or as an alternative to literal infringement. The question then becomes what is required to adequately plead a DOE claim.
Continue Reading

Patent Venue for Method Claims: The "One-Step Rule" at a Crossroads
Jaemin Sung*, Lauren Dreyer
Comcast has asked the Federal Circuit to weigh in on how patent venue applies to method claims. The company's recent mandamus petition challenges how the Eastern District of Texas applied the patent venue statute, 28 U.S.C. § 1400(b), and could reshape where tech companies face suit when their products rely on distributed software or cloud-based systems.
*Jaemin Sung, a Law Clerk at Baker Botts, assisted in the preparation of this article.
Continue Reading

South Korean AI Framework Implementation Package: On September 8, 2025, South Korea's Ministry of Science and ICT released a comprehensive draft package of regulations to implement the country's AI Framework Act, which takes effect January 22, 2026. The package includes an Enforcement Decree, Safety Assurance Notice, and Business Responsibilities Notice that establish compute-based thresholds (generally 10²⁶ FLOPs) and risk-based criteria for determining which AI systems must implement lifecycle safety measures and reporting obligations. Operators of "high-impact AI" must establish documented risk management programs, prepare explanation plans, institute user-protection measures, ensure human oversight, and publicly post core governance information while maintaining records for five years. The framework creates a layered regulatory structure that aligns Korean AI governance with international approaches like the NIST AI RMF and EU AI Act. Nick Palmieri writes about this development here: "South Korean Ministry of Science and ICT Issues Package of Regulations to Supplement AI Framework Act."

Supreme Court Petition on AI-Generated Works: Dr. Stephen Thaler filed a petition for certiorari with the U.S. Supreme Court on October 9, 2025, asking whether AI-generated works without direct human authorship can receive copyright protection. The petition follows successive denials by the Copyright Office, D.C. District Court, and D.C. Circuit Court, all of which held that purely AI-generated works lack the requisite "traditional human authorship." Thaler argues that the Copyright Act's text contains no human-authorship requirement, that the work-for-hire framework already contemplates nonhuman authors, and that excluding AI-generated works undermines copyright's constitutional purpose to promote progress. If the Court grants review, the decision could establish fundamental ground rules for copyright protection of AI-generated content at a critical moment in the technology's development. You can read more about this petition here: "AI As An Author - 'Creativity Machine' Creator Asks Supreme Court to Reconsider Authorship Requirements."

Patent Eligibility Win for AI Training Methods: The USPTO's Appeals Review Panel vacated a Patent Trial and Appeal Board Section 101 rejection on September 26, 2025, in a case involving DeepMind's machine learning patent application. The Board had found claims directed to training methods for continual learning to be abstract mathematical calculations, but the ARP held that the claims as a whole integrate the abstract idea into a practical application by addressing "catastrophic forgetting" and improving AI system operation through reduced storage requirements and preserved performance across tasks. The decision affirms that AI and machine learning inventions providing specific technical improvements can satisfy Section 101 when integrated into practical applications, and cautions against overbroad application of eligibility rejections when Sections 102, 103, and 112 are the proper tools for limiting patent scope. Nick Palmieri writes about this decision here: "USPTO Appeals Review Panel Vacates PTAB Section 101 Rejection."

Senate AI Chip Priority Legislation: The Senate's fiscal year 2026 National Defense Authorization Act, passed on October 9, includes the GAIN AI Act, which would require semiconductor manufacturers seeking export licenses for advanced integrated circuits to certify that U.S. customers have been offered right-of-first-refusal, receive pricing terms at least as favorable as foreign customers, and that no domestic backlog exists or will be created by the export. The provision aims to address procurement delays for AI chips amid supply constraints, with proponents arguing this protects economic competitiveness and national security. However, passage is uncertain as the House NDAA lacks a similar provision, leaving the Act's fate to the conference committee reconciliation process. Jason Wilcox writes about this development here: "'America First' for AI Chips: Senate Bill Seeks to Reshape Semiconductor Market."

October 2025 Intellectual Property Report Recap
In case you missed it, here is a link to our October 2025 Intellectual Property Report.

The content of this article is intended to provide a general guide to the subject matter. Specialist advice should be sought about your specific circumstances.