Attributional Ambiguity in AI-Mediated Harm: How Perceived AI System Limitations Shape Forgiveness and Regulatory Demand

Artificial intelligence (AI) increasingly shapes corporate decision-making, but AI-mediated harms, such as AI-generated discrimination, complicate how individuals assign responsibility and respond. Grounded in attribution theory, this study examines how crisis responsibility perceptions (intentionality, accountability, and internal locus of causality as corporate negligence) lead to anger and empathy, which predict organization-focused forgiveness and system-focused demand for regulatory intervention. Conducting an online survey (N = 705), the findings demonstrate a dual pathway in which publics simultaneously pursue relational repair with the offending organization and systemic governance of AI. This further highlights the role of attributional ambiguity in shaping divergent public responses to AI-mediated corporate crises.