This decision concerns an invention that relates to an access-control method for a hearing instrument.
Since the distinguishing features were considered non-technical, the EPO refused to grant a patent. Here are the practical takeaways from the decision T 0521/23 (Access-control method for a hearing instrument / SONOVA) July 7th, 2025, of the Technical Board of Appeal 3.5.04.
Key takeaways
The invention
1.1 The opposed patent concerns a method of controlling access to services provided by a hearing instrument, such as volume control or allowing fitting parameters to be read from and written to the hearing instrument.
The invention addresses safety risks in connection with connecting hearing instruments to external client devices:
1.2 According to the patent, modern hearing instruments can connect to various external client devices, such as smartphones or fitting stations. As set out in paragraph [0003] of the patent specification, a problem with prior-art systems is said to be that, once connected, a client’s applications may have full access to the hearing instrument. This is described as posing a safety risk, as it grants even non-trusted applications the ability to modify safety-critical hearing-instrument configurations.
Fig. 8 of EP 3 542 554 A1
-
Claim 1 of the main request
Is it technical?
D2 is considered as a suitable starting point for assessing inventive step. Featues (a) to (c) are seen as disclosed in D2, this is not contested. The Board makes the following observations regarding the technicalitiy of features (d) to (f):
2.2.1 The respondent submitted that the technical effect to be attributed to the distinguishing features was a “simplification” of the access-control scheme, leading to “low resource requirements”. In support of this, it argued that the skilled reader would always understand the invention in the context of a “hearing instrument”, which was, in the respondent’s opinion, an inherently resource-constrained device, a fact that it considered to be confirmed by document D2 itself (see D2, paragraph [0004]). This echoes Reasons 4.2 of the appealed decision, in which the opposition division, referring to paragraph [0015] of the opposed patent, found the technical effect of the distinguishing features to be “client specific service access requiring low resources”. In that regard, the respondent argued that the claimed method, which relied on a single comparison against a client’s “highest security level” as per feature (d), was inherently simpler than checking a detailed list or matrix of rights for each client, the latter being a method it considered to be representative of the skilled person’s common general knowledge as exemplified in document D8. During the oral proceedings before the board, the respondent attempted to concretise this by giving an example of managing several thousands of hearing-care professionals (HCPs) via a simple classification scheme where all HCPs share the same security level.
2.2.2 The board finds these arguments generally unconvincing. These arguments rely on reading limitations into a claim which are not supported by its wording. They require the skilled reader of present claim 1 to understand the claim in the specific context of a resource-limited device and to assume that the claimed architecture is necessarily implemented in its simplest form. In view of the wording of feature (b), the skilled reader of claim 1 could only speculate as to whether the actual “service” requested by the “client” (being relevant to the performance of method step (d)) is notified by the “client authenticator”, as advocated by the respondent, or by another, yet undefined, data message. In addition, it is not derivable from claim 1 whether, in feature (e), the “security level” to be selected (arguably by an administrator according to the respondent) from many security levels is supposed to refer to the “security level associated with the service request” or to a “security level assigned to the client”, so that the skilled reader cannot even establish whether the “client” is necessarily associated with more than one security level (i.e. “highest security level”).
However, the assessment of inventive step must be based on the whole scope claimed, not on a preferred or commercially advantageous embodiment. The board notes the respondent’s reference to the “business success of respective products” but considers this to be irrelevant to the assessment of inventive step, particularly where a credible technical effect is not established. The respondent’s further example of simplifying the management of several thousands of HCPs by assigning them all to a single security level is equally unpersuasive. As the appellant correctly observed, claim 1 as granted does not relate to such a grouping or classification of clients: this is merely another example of an argument based on a specific, undisclosed implementation rather than on the technical features of the claim itself.
2.2.3 More specifically, the board cannot acknowledge the alleged technical effects of “simplification” or “low resource requirements” as being credibly achieved over the entire scope of claim 1. This is because claim 1 is silent about the resource requirements of the claimed method. It limits neither the number of security levels in the “plurality”, nor the frequency with which the comparison steps are actually conducted. Hence, an implementation with, for example, a complex hierarchy or involving frequent access requests could in fact be resource-intensive. Furthermore, claim 1 as granted does not preclude the possibility of “offloading” computationally intensive tasks to a more powerful peripheral device, such as a smartphone. For this reason, the respondent’s argument that “the skilled reader would not consider complex algorithms when dealing with a hearing instrument” must be discarded.
2.2.4 In view of the above, the board cannot derive any credible technical effect of “simplification” or “reduced resource usage” from the distinguishing features (d) to (f). At most, the effect of these features can be seen as the definition of a particular access-control policy (see also the description as filed, e.g. page 5, line 20 to page 6, line 8). This policy dictates that access is granted if a “requested security level” is below or equal to the “highest security level assigned to the client”. The board concurs with the appellant’s argument, advanced during the oral proceedings before the board, that such a policy is primarily administrative in nature and that its advantages, if any, consequently lie on the administrative rather than the technical side.
2.2.5 Based on the effect identified in point 2.2.4 above and the well-established COMVIK approach (cf. T 641/00, headnote II), the objective technical problem consists therefore, at most, in the technical implementation of this specific, non-technical security policy on the hearing instrument disclosed in document D2.
The Board is further of the opinion that the skilled person would have arrived at the solution:
2.3.2 The board considers that the skilled person, starting from D2 and seeking to implement the security policy defined in the objective technical problem (see point 2.2.5 above), would have been motivated to use a hierarchical model. The respondent’s argument that the common general knowledge pointed only towards a complex matrix is unconvincing given the ubiquity of hierarchical access-control models. The board further concurs with the appellant that document D2 itself, in paragraph [0038], hints at a multi-level security architecture by disclosing a basic connection authenticated by a pre-established link (such as Bluetooth**(®) pairing) on top of which a further service is authenticated by a digital signature. The respondent’s counter-argument that this merely discloses two sequential authentication steps and not a hierarchy is not accepted. These are two distinct methods for accessing services, implying, typically, different levels of trust and security.
2.3.3 The respondent’s main defence, namely that the skilled person would have been deterred from applying a concept from the field of computer science due to the resource constraints of a hearing instrument, may have convinced the opposition division (cf. Reasons 4.2 of the appealed decision, last sentence) but in the board’s view is without merit. As established in point 2.2.3 above, the claimed method could involve “outsourcing” in the sense that it is at least partially “offloaded” to a more powerful peripheral device: it is not required to be executed solely on a hearing instrument. Even if it were, the board considers that the skilled person would have employed obvious strategies to mitigate resource limitations, such as optimising the security features to minimise their resource footprint or tailoring these features to the specific needs of the hearing device and its user.
The respondent argued that “outsourcing” tasks to a connected device would have led the skilled person away from the claimed invention. This, however, was a mere allegation for which no substantive reasoning was provided. Moreover, the argument misconstrues the board’s reasoning: the board did not suggest “outsourcing” as a step towards the claimed invention, but rather as an obvious alternative strategy available to the skilled person at the relevant date for mitigating resource constraints. The existence of such an obvious alternative undermines the respondent’s argument that their specific on-device solution required an inventive step. The respondent’s own argument that only the most critical steps (e.g. those underlying features (d) and (f)) must be carried out on the hearing instrument itself implicitly acknowledges that other steps can be externalised, further supporting the board’s view.
2.3.4 The board therefore holds that the solution defined by features (d) to (f) would have been an obvious course of action for the skilled person based on the disclosure of document D2 in combination with their common general knowledge.
As a result, the Board came to the conclusion that, in the absense of a specific use of the obtained data, the claimed subject-matter cannot be considered inventive. The same applies to the remaining requests. Thus, the appeal was dismissed.
More information
The decision can be found here: Here are the practical takeaways from the decision T 0521/23 (Access-control method for a hearing instrument / SONOVA) July 7th, 2025, of the Technical Board of Appeal 3.5.04.