Recent T decision demonstrates the EPO’s position on technical character and sufficiency in relation to AI

Continuing our series on biotechnology and artificial intelligence (AI), here we provide an update on a recent European Patent Office (EPO) Board of Appeal Decision relating to AI, specifically a method for analysing image data from cells growing in culture.

Earlier this year, the Boards of Appeal of the EPO rendered the T 0660/22 decision, examining a patent application jointly filed by The University of Tokyo and Nikon Corporation. The decision is particularly interesting as the findings of the Appeal represent an implementation of the Enlarged Board of Appeal G1/19 decision, summarised by Venner Shipley here and here. In G1/19, the Enlarged Board considered that calculated numerical data can have technical character if it has an inherent technical quality or if it serves a technical purpose. Measuring physical quantities, whether directly or indirectly, can also be considered to have technical character.

The T 0660/22 Appeal was against a decision of the Examining Division to refuse the application for lack of compliance with Article 83 EPC. Under Article 83 of the European Patent Convention (EPC), a patent application needs to present a detailed, reproducible disclosure across the full breadth of the claimed subject matter.

The application in question sought protection for an analysis device and a method for processing sequences of microscope images to characterise cells. The claims related to a cell region separation unit, a characteristic quantity extraction unit, and a mechanism analysis unit (see claim 1). Together, these components aimed to calculate correlations among characteristic qualities associated with various cellular elements (see for example paragraphs 73, 75, 77, 79, 82, 84 of the application). The elements and their correlation strengths were visualised as a graphical model, in which the elements form the nodes and each pair of elements are linked by a line with a thickness representing the correlation strength (see paragraphs 125 and 126 and figures 23 to 25). This offered a novel approach to cell analysis that could, for instance, enhance drug screening efforts by better quantifying cellular responses.

The Board’s assessment primarily focused on the sufficiency of the disclosure and technical character. The Examining Division argued that the claims covered the computation of a “practically unlimited” number of characteristic quantities and corresponding first and second correlations and that it was not possible to derive which mechanisms were sought and which vital phenomena needed to be observed. The Examining Division also stated “in the present case, there is no technical purpose or an associated technical effect because the specific computations and the creation of the model illustrating relationships between predetermined elements of the cell is merely a way of manipulating data. The outcome, in the form of an interpretation of the vital phenomena relating to cells, as broadly stated in the description of the application, is left at the latitude of the user (in this case a biologist)”.

The Appeal board agrees with the Examining Division and concluded that the claims lacked sufficiency. The Appeal Board also found that claimed aspects—particularly those related to the calculation of correlations—failed to establish a technical contribution over the prior art. The reasoning centred on whether an indirect measurement method provided a technical contribution.

The Board considered that “any measurement method within the meaning of G 1/19 (reasons 99), whether indirect or not, or corresponding device, must be intended to determine a specific and predefined physical quantity. A measurement method is considered “indirect” when it determines the desired physical quantity by measuring one or more different physical quantities and applying a known factual relationship between them”.

The Board considered that unless a factual relation­ship between physical quantities is actually established, these values are “just data, which may be used, for example, to gain scientific knowledge” (G 1/19, reasons 98) and obtaining them is not an indirect measurement within the meaning of G 1/19, reasons 99. Consequently, the Board concluded that the device and method did not provide indirect measurements in the sense of G 1/19, reasons 99, nor a technical contribution, and so were not considered to involve an inventive step in the sense of Article 56 EPC.

The Board also considered an auxiliary request claiming a cell manufacturing method, but it was concluded that the request was too broadly claimed and did not comply with the provisions of Article 83 EPC.

This decision reflects a broader challenge in patenting AI-driven and software-based innovations: simply correlating numerical data or applying digital analysis techniques does not automatically satisfy the technical requirement unless they can be linked to specific physical measurements or technical purposes. In particular, T 0660/22 continues to support the conclusions of G1/19, because it reinforces the principle that a patentable invention must deliver a concrete technical contribution beyond its mere computer-implemented steps.

The decision further invites a broader discussion on defining what constitutes an indirect measurement and inherent technical quality. This is particularly important when the claimed innovation spans AI, image analysis, and cell biology areas where the boundaries between analysis and technical solutions can become blurred. As AI-driven biotechnology continues to evolve, ensuring alignment with established EPO requirements will remain key to navigating the increasingly complex patent landscape.

If you have any questions on this, please reach out to Sophie Newgas or a member of the team.