Preventing promotion-focused goals: The impact of regulatory focus on responsible AI

Abstract

Implementing black-box artificial intelligence (AI) often requires evaluating trade-offs related to responsible AI (RAI) (e.g., the trade-off between performance and features regarding AI's fairness or explainability). Synthesizing theories on regulatory focus and cognitive dissonance, we develop and test a model describing how organizational goals impact the dynamics of AI-based unethical pro-organizational behavior (UPB). First, we show that promotion-focused goals increase AI-based UPB and that RAI values act as a novel mediator. Promotion-focus goals significantly lower fairness in Study 1A and explainability in Study 1B, mediating the relationship between regulatory focus and AI-based UPB. Study 2A further supports RAI values as the driving mechanism of AI-based UPB using a moderation-by-processes design experiment. Study 2B provides evidence that AI-based UPB decisions can, in turn, lead to more unethical RAI values for promotion-focused firms, creating a negative RAI feedback loop within organizations. Our research provides theoretical implications and actionable insights for researchers, organizations, and policymakers seeking to improve the responsible use of AI.