The aim of this study is to investigate the Critical Friend (CF) and socio-moral decision-making, by understanding the CF in a formal way. CF, utilized to be understood as characterized behavior, is in contrast with state-of-the-art AI-systems argued to be a responsible construct, suitable for characterizing such formalization. This is done by asking 1) How to characterize a formal definition of the CF?, and 2) How to assess socio-moral decision-making of the CF? Results are a formal framework (in terms of an ontology and a transition system) of the CF, and an initial assessment of socio-moral decision-making of the CF. Results can be utilized by researchers to better understand socio-moral decision-making in AI-systems.