This paper defines the structural configuration of responsibility in AI-Scored Soci-eties. An AI-Scored Society (AISS) refers to a social environment in which AI-mediatedvisibility structures influence human behavior through the continuous allocation of at-tention, credibility, and opportunity. Within such environments, artificial intelligencedoes not function as an autonomous decision-making subject. Instead, AI operates asan infrastructural environment that structures the conditions under which decisionsappear to occur.Under these conditions, responsibility cannot be attributed to a single actor. Rather,responsibility becomes distributed across states, corporations, algorithmic infrastruc-tures, and human agents that collectively shape AI-mediated decision environments.This paper introduces the concept of responsibility allocation to describe the struc-tural distribution of responsibility across these institutional actors. It further proposesthe concept of responsibility drift, referring to the tendency for responsibility to dis-perse, shift, and accumulate within complex AI-mediated governance environments.The objective of this paper is not to propose policy solutions or normative frame-works, but to conceptually define the responsibility structure emerging in AI-ScoredSocieties.
Kawazoe Tsutomu (Fri,) studied this question.