Virtual affective consciousness and raw social AI
In this paper, we attempt to philosophically eliminate the conceptual problems in develop-ing affective artificial agents, and then put forward an industry and research direction called Raw Social AI. We first provide a set of laws for affective concepts understanding, and then suggest that we should build unconscious concepts with uncertainty. We address the problem of qualia supply, and discuss how to implement the initiative of affective agents. We suggest that initiative be the cue of utility of affective agents. We advocate that the initiative of affective agents must be realized as social services. Raw social AI is about using AI to record, model and predict people's emotional reactions in the real or virtual world. It is necessary for us to establish the infrastructure of raw social AI.