Towards Patient-oriented Transparency

The paper introduces a patient-oriented concept of transparency in application to medical AI systems. It is argued that the patient’s perspective differs significantly from the user’s perspective and that there is a need to re- think what transparency entails for a patient rather than for a professional user. It is suggested that there are two major factors that influence the concept of transparency if we take the patient into account: (a) the active patient paradigm and (2) the patient’s trust in the medical sphere. It is further argued that, for an active patient, it is important to be able to make an informed decision about her health and voice criticism of a diagnosis/treatment, if she has sufficient ground for it. This poses one type of constraint on the concept of transparency. The other type of constraint derives from the fact that, in order to be trusted, medical expertise must be rooted in scientific methodology.