Effects of AI versus human source attribution on trust and forgiveness in the identical corporate apology statement for a data breach scandal

This study investigates the effects of AI versus human source attribution on trust and forgiveness in the identical AI-generated corporate apology statement for a simulated data breach scandal. While AI-generated messages hold promise in crisis communication, their impact on public perception remains understudied. The research was inspired by incidents where ChatGPT was used to generate official apology statements, raising questions about the authenticity of AI-generated apologies. Using a fictitious retail company's apology statement, crafted with the assistance of ChatGPT, participants were randomly assigned to conditions indicating the statement was AI-aided, human-written, or unspecified (control). The results indicate that participants attributed higher levels of forgiveness intention and trust to the statement credited to humans compared to AI-generated statements. Additionally, the human-attributed statement was perceived as more empathetic and sincere than the AI-attributed statement. Mediation analysis results revealed that empathy mediated forgiveness intention and trust in human-authored statements, while perceived sincerity mediated these factors in AI-aided statements. These findings suggest that source attribution significantly influences public perception of organizational apologies during crises. This study contributes to understanding the evolving role of AI in crisis management and underscores the importance of ethical and transparent communication practices.