Unpacking Generative Artificial Intelligence-Powered Crisis Communication From a Moral Angle: The Role of Moral Outrage, Authenticity, and Forgiveness

Abstract

This study aims to advance crisis communication by incorporating morality in understanding crisis communication and explicating effective strategies for using artificial intelligence (AI) technology (e.g., ChatGPT) in the context of the ‘preventable’ cluster of crisis types. To this end, we conducted an online experiment, adopting a 3 (crisis type) × 2 (crisis response source label) between-subjects design. The results indicate that scansis elicited greater moral outrage than other crisis types and that a crisis response from ChatGPT led to lower authenticity than a crisis response from a PR representative. Indirect effects analyses demonstrated that moral outrage, authenticity and forgiveness mediated the effects of crisis type and message source label on behavioural outcomes. Theoretical and practical implications are discussed.