Socio-Affective Interpretation in AI Discourse: A Descriptive Frame

Abstract Public and institutional discussions of AI systems often involve disagreement that persists even when technical specifications, empirical performance measures, and organizational arrangements are shared. In addition to technical explanations, participants frequently rely on expectations, legitimacy judgments, and affective attributions when interpreting AI behavior and impact. This paper presents a bounded, descriptive framework for examining these patterns as socio-affective interpretation. The analysis treats socio-affective dynamics as interpretive structures rather than as evidence of system failure, governance absence, or design deficiency. It identifies recurring dimensions along which interpretation varies, including attribution, legibility, authority projection, affective salience, and frame compatibility. It also specifies conditions under which interpretive disagreement persists without convergence. These dynamics are examined as features of discourse and meaning-making, not as issues requiring resolution. The framework is intended to support conceptual clarification in interdisciplinary research settings where multiple explanatory frames coexist and disagreement arises independently of contested empirical evidence. It does not advance normative claims, policy recommendations, governance mechanisms, or design interventions. Its contribution lies in clarifying how interpretive disagreement emerges and persists in AI discourse, and in outlining analytical boundaries that distinguish interpretation from evaluation, diagnosis, or prescription.