Critical infrastructures (CIs) embody cyber-physical-social systems (CPSSs) where physical entities are integrated with cyber components, shaping service delivery through end-user behavior. The seamless operation of CIs is vital for society, and the CPSS resilience relies on interdependencies with AI-integrated technologies. The complexity of the system, and the interconnections with other infrastructures, along with the socio-technical transition towards digitization raised the necessity of implementing Resilience Engineering. This motivates exploration of the scientific literature on resilience key performance indicators (R-KPIs) which support strategies for ensuring service continuity. Therefore, this article aims to identify R-KPIs for AI-integrated CIs and prioritize the extracted R-KPIs using a hybrid Multi-Criteria Decision-Making (MCDM) approach. The results show the importance of employing R-KPIs that measure risk probability, energy self-sufficiency level of the system under study, and performance indicators including functionality loss, recovery time, and minimum performance level after disturbance as the most effective R-KPIs in the domain of this study. After identifying and prioritizing the R-KPIs, a general framework is proposed to employ these R-KPIs in modeling the resilience of a CPS. Finally, a case study demonstrates the implementation of the framework and KPIs in a real-life scenario.
