What happened
The Patent Trial and Appeal Board (PTAB) is clarifying patent disclosure requirements for artificial intelligence inventions under 35 U.S.C. § 112(a). AI's inherent opacity, training data dependency, and non-deterministic behaviour challenge traditional patent law's demands for written description and enablement. In Ex Parte Kirti (2021), PTAB found a machine learning model patent sufficient, citing disclosed model type, training methodology, inputs, and outputs. Conversely, Ex Parte Allen (2021) affirmed rejection for a natural language processing method, lacking algorithm specifics or scoring details.
Why it matters
Securing intellectual property for AI innovations now requires explicit disclosure of model types, training methodologies, and data characteristics. For founders and CTOs, patent applications must detail how AI systems function, not just their outcomes. PTAB's rulings in Kirti and Allen establish that general descriptions are insufficient; specific algorithms and scoring mechanisms are necessary unless the underlying machine learning approach is conventional. This shifts the burden to articulate AI's "how," impacting investment and competitive advantage.
Subscribe for Weekly Updates
Stay ahead with our weekly AI and tech briefings, delivered every Tuesday.
