argument: Notizie/News - Intellectual Property Law
Source: Kluwer Patent Blog
AI and computer-implemented patent claims face increasingly similar assessments of technical contribution in the UK, UPC and EPO, argues Eden Winlow of Bristows. Winlow distinguishes that convergence from full harmonisation because the systems retain different approaches to inventive step. According to the analysis, the UK Supreme Court’s decision in Emotional Perception AI Ltd v Comptroller-General of Patents displaced the Aerotel framework in favour of principles drawn from Duns, COMVIK and G 1/19. The UKIPO subsequently issued guidance addressing the revised approach during search and examination. A reference to technical means, such as computer hardware, can ordinarily satisfy the initial requirement for an invention, but does not establish patentability. The subsequent assessment identifies features capable of supporting inventive step through their contribution to the invention’s technical character, including contributions arising from interaction with other features.
On remittal in Emotional Perception, a UKIPO hearing officer refused an application concerning a neural network trained to recommend files through semantic similarity derived from physical properties. The hearing officer found that improved semantic recommendations were not a technical result; after the associated training features were excluded from consideration, the remaining subject matter was obvious over common general knowledge. In FLIR Belgium BV, the hearing officer accepted that some virtual-gauge features contributed technically through their operation within the system, excluded a panning feature, and returned the application for further examination. Winlow also examines Abbott v Sinocare, where the UPC Court of Appeal considered event icons alongside glucose data to contribute to improved human-machine interaction supporting diabetes management. Winlow stresses that Abbott arose in preliminary injunction proceedings and provides a limited illustration, rather than a complete method for assessing complex AI claims. The commentary concludes that technical contribution should be assessed transparently before obviousness, while acknowledging unresolved questions about applying the revised UK approach and the UPC’s developing methodology.