Abstract: With the rapid expansion of the artificial intelligence (AI) industry, a wide array of AI services has emerged to meet diverse needs. However, research on the relationship between AI services and consumer moral behavior remains in its early stages, not only limited in scope but also facing three primary shortcomings. First, existing studies tend to focus on specific types of moral or immoral behaviors. For instance, some studies examine how AI contexts may prevent consumers from engaging in dishonest behaviors, such as lying (Anthony & Cowley, 2012) or purchasing pirated goods (Kos Koklic et al., 2016), as well as how AI may encourage moral behaviors like donating to charity (Dunn et al., 2020) or volunteering (Macdonnell & White, 2015). However, there is a lack of systematic differentiation and discussion of the dual nature of consumer moral behaviors, encompassing both moral/immoral actions and "doing good" versus “avoiding harm.”
Second, current research primarily highlights the negative effects of AI services on consumer moral behavior. For example, when AI (rather than human) cashiers in supermarkets undercharge customers, consumers tend to feel less guilt and, consequently, are less inclined to correct the cashier's mistake (Giroux et al., 2022). Similarly, when AI replaces human spokespersons in charitable projects, AI presenters tend to trigger more rational, utilitarian judgments in consumers, thereby reducing their donation amounts (Zhou et al., 2022). However, this focus on the adverse impacts of AI has led to an oversight of the potentially dual-edged sword effects that AI services may exert on consumer moral behavior and the mechanisms underlying these effects.
Third, existing literature has yet to investigate the boundary conditions that influence the effects of AI services on both types of consumer moral behavior. This gap in understanding not only restricts a fuller demonstration of the complexity and variability of AI’s impact on consumer morality but also limits practical guidance for businesses and society regarding the moral use of AI.
In response to these research gaps, this study proposes an innovative approach from three perspectives. First, by introducing the moral duality theory from moral psychology into the field of AI ethics, this study categorizes moral behaviors into two types: proscriptive moral behaviors (“doing good”) and prescriptive moral behaviors (“avoiding harm”) (Janoff-Bulman et al., 2009). This framework allows a systematic distinction between the unique impacts of AI services on these two types of moral behaviors, providing a more nuanced understanding of AI's influence.
Second, grounded in the double-process theory of moral judgment in moral psychology—which asserts that moral behavior formation involves both moral emotions and moral cognition (Greene et al., 2001; Greene et al., 2004; Greene, 2009)—this study aims to reveal the dual-edged effects of AI on moral behavior and their underlying mechanisms by exploring both cognitive and emotional aspects. By doing so, it examines how different AI services may either promote or inhibit moral behavior based on these two psychological processes.
Third, this study thoroughly identifies the boundary conditions of AI's impact on consumer moral behavior by examining factors related to the AI itself, the consumers, and the types of moral behavior in question. Additionally, it seeks to uncover the moderating factors that exert varying effects on the dual aspects of moral behavior, providing a more comprehensive understanding of how AI services interact with consumer morality.
In conclusion, by integrating perspectives from moral duality theory and the double-process theory of moral judgment in moral psychology, this research is the first to systematically investigate the mechanisms and boundary conditions of AI's impact on consumer moral behavior. This study not only contributes theoretical insights but also offers practical guidance for enhancing consumer moral consciousness, helping businesses, and aiding public sectors in designing strategies to promote moral behavior through AI innovations.

Key words: artificial intelligence service, moral behaviors, double-edged sword effect, moral duality theory, double-process theory of moral judgment