•AI in elderly care promises to be predictive, personalized, preventive, participatory.•AI could also bring disruption of care processes and relationships.•Model to frame ethical risks is lacking.•4d-model: risk of depersonalized, discriminatory, dehumanized, disciplining care.•Joint effort of all stakeholders to prevent 4d-risks needed. Gerontechnology based on Artificial Intelligence (AI) is expected to fulfill the promise of the so-called 4p-medicine and enable a predictive, personalized, preventive, and participatory elderly care. Although empirical evidence shows positive health outcomes, commentators are concerned that AI-based gerontechnology could bring along the disruption of elderly care. A systematic conceptualization of these concerns is lacking. In this paper, such a conceptualization is suggested by analyzing the risks of AI in elderly care as “4d-risks”: the depersonalization of care through algorithm-based standardization, the discrimination of minority groups through generalization, the dehumanization of the care relationship through automatization, and the disciplination of users through monitoring and surveillance. Based on the 4d-model, strategies for a patient-centered AI in elderly care are outlined. Whether AI-based gerontechnology will actualize the 4p-perspective or bring about the 4d-scenario depends on whether joint efforts of users, caregivers, care providers, engineers, and policy makers will be made.