Enlarged Board of Appeal of European Patent Office clarifies examination of patentability of computer-implemented simulations
The Enlarged Board of Appeal of the European Patent Office (EPO) has concluded that the long-established “COMVIK” approach for computer-implemented inventions applies also to computer-implemented simulations. Accordingly, inventions directed to computer-implemented simulations are expected to be examined by the EPO as any computer-implemented invention and to be subject to the same two hurdles set by the COMVIK approach for assessing compliance of such an invention with the eligibility requirement (first hurdle) and the inventive step requirement (second hurdle).
The Enlarged Board of Appeal (EBoA) of the European Patent Office (EPO) issued decision G 1/19 concluding that the long-established “COMVIK” approach for computer-implemented (CI) inventions (T 641/00) applies also to computer-implemented simulations. Accordingly, computer-implemented simulations are expected to be treated by the European Patent Office (EPO) as any computer-implemented invention and to be subject to the same two hurdles set by the COMVIK approach to assess compliance of such an invention with the eligibility requirement (first hurdle) and the inventive step requirement (second hurdle).
The case considered by decision G1/19 relates to the computer simulation of a crowd of pedestrians in an environment, and usable as part of a process for designing places such as a stadium or a train station. The first-instance Board was tempted to consider that the simulation did not produce a technical effect, for “lack of direct link with physical reality”, as opposed to decision T 1227/05 relating to an electric circuit simulation. In decision T 1227/05, the Board had deemed that the simulation of a circuit subject to 1/f noise constituted an adequately defined technical purpose for a CI-simulation.
Decision G1/19 has recognized that computer-implemented simulations may be technical, and thus eligible for patent protection. In this respect, the decision is particularly welcome, since inventions in the field of machine learning (ML) and artificial intelligence (AI) are on the rise and find applications in many areas, such as, for example, drugs, medical diagnosis, chemical processing, image and speech processing, transport and communication networks. Building models is at the core of AI and ML. The decision may thus have far-reaching implications. For example, depending on the requirements set on the link between the simulation and the physical reality, an applicant for a European patent may find itself in a very different situation depending on whether its business is focused on the manufacture of the product designed using the simulation, and/or the designing of the software.
However, decision G1/19 does not appear to provide any additional clear guidance to applicants of European patent applications, apart from the guidance already available under the COMVIK approach, for meeting the inventive step requirement.
According to the COMVIKapproach, the first hurdle is relatively easy to pass because a computer is considered to be a technical means, and the use of a technical means is deemed to be enough to establish the presence of a technical character and thus eligibility under Article 52(2)(3) EPC. Therefore, eligibility before the EPO, unlike before the U.S. Patent and Trademark Office, may be obtained by simply directing a claim to “A computer-implemented method” (emphasis added).
The second hurdle, the one relating to the inventive step requirement, however, is far from being easy to pass for a computer-implemented invention, which generally consists of a mixture of technical and non-technical features, especially whenever the computer is the only technical feature. All non-technical features of a claim, in fact, such as formulae or program steps, that do not contribute to the technical character, and thus to the solution of a technical problem by providing a technical effect, are not taken into account when assessing inventive step. The only case where non-technical features may be considered is if they contribute to produce a further technical effect,i.e., an effect going beyond the mere implementation of the simulation on a computer, beyond normal interactions with the computer.
Thus, based on the established case law, recognized by the EBoA in decision G1/19, a claim may be considered to involve inventive step if, for example, it involves the use of a technical input, such as a measurement of a physical value, or a technical output, such as a control signal or steps for physically controlling a machine, or an adaptation to the computer or its operation. However, the EBoA also appears to leave some potential room for patentability to other types of claims, without, unfortunately, providing a specific guidance thereto.
For example, the EBoA states that the mere fact that a simulation is based on technical principles underlying the simulated system or process is neither sufficient nor necessary to confer inventive step and thus patentability to the invention. Instead, the EBoA seems to indicate that the intended technical purpose of the simulation needs to be at least implied in the claims (see points 124, 131). The EBoA deems that the existence of technical principles are not sufficient for the simulation to be considered as “solving a technical problem”, since otherwise, “CI-simulations would hold a privileged position within the wider group of CIIs without there being any legal basis for such a privilege” (section 141).
Conversely, a “direct link with physical reality” is not required in all cases, if only because the notion of technicality needs to remain open: additional technical effects can for example occur within the computer (by specific adaptations of a computer or a data transfer). For a C-I simulation to be patentable, it has to provide a further technical use that has an impact on physical reality, with such further use having to be at least implicitly specified in the claim. While an additional technical effect linked to the inputor the output may confer inventive step, in the EBoA’s view, most C-I simulations are unlikely to be patentable as “most simulations as such” may have few technical effects as far as input and output are concerned, unless the intended technical purpose of the simulation giving rise to the alleged technical effect is at least implied in the claims.
What may be considered “technical” for the purpose of assessing inventive step is at the heart of decision G1/19. The EBoA puts a strong emphasis on keeping the definition of “technical” open and examining simulation inventions on a case-by-case basis. According to an old definition, “a technical system or process implies that an object is created or a process is run with some purpose based on human creativity”, be it a “potential”, a “virtual”, or a “tangible” effect.
One way to make simulations sufficiently technical could be to show an increased accuracy in representing reality. Whether a simulation contributes to the technical character of the invention does not depend on the quality of the model or the degree to which the simulation represents reality. Instead, the Board holds that the accuracyof a simulation is a factor that may have an influence on a technical effect going beyond the simulation's implementation and may therefore be taken into consideration in the assessment of inventive step.
The underlying models and algorithms may contribute to the technical character when, for example, they constitute a reason for adapting the computer or its operation or if they are the basis of a technical use associated with the result of the simulation, provided that such adaptations or uses are at least implicitly specified in the claim.
In this context, a simulation aiming merely at forecasting the weather stands no chance, independently from the number of parameters used as inputs, as opposed to the same simulation aiming at forecasting the weather in order to control a store.
Before the issuance of decision G1/19, people were concerned, regarding simulations in particular, about the fact that it appears that the technical effect is, in a large number of cases concerning these inventions, only simulated on a computer, in silico, and hence would equate to a non-technical “mental act”, thus resulting in a non-eligible invention, or at least to an effect that would not be taken into account when assessing inventive step. These concerns, in the light of decision G1/19, may be now considered unjustified. The difficulty to comply with the inventive step requirement, however, remains intact.
Indeed, as discussed above, there will be no special treatment for simulations as compared to any other computer-implemented invention. This is no revolution or case law reversal, since decisions on simulations as such are rare (applicants would eventually introduce an unambiguous technical step). Applying the COMVIK approach to simulations seems to be reasonable since they are examples of methods implemented by computers.
The decision will likely make life more complicated for applicants attempting to rely merely on, for example, a physical model underlying the invention, on the physical nature of the parameters taken into account or modelled, or on the potential physical use of the simulation’s results as features contributing to the technical character of the simulation. These do not appear to be viable routes.
The decision has established that C-I simulations may be technical. However, what still remains vague is the “further technical use” of the outcomes of the simulation that should be specified at least implicitly to contribute to the technical character of the invention for the purpose of establishing inventive step. This aspect needs to be closely monitored in the future.