UC Irvine

Consumers increasingly receive marketing content and services from artificial intelligence (AI)-based sources, including LLM chatbots, AI-powered service apps, and customer service AI agents. Merely learning that a service or product is AI- (vs. human)-based often results in relatively less favorable responses from consumers. Beyond this evaluative bias, I argue that consumers also become more cautious and exhibit heightened avoidance of negative outcomes when AI is involved. Drawing on regulatory focus theory, I propose that AI can function as an antecedent to consumers’ regulatory focus: AI- (vs. human)-based offerings induce a relatively greater cautious mindset (prevention focus) and a lower eager mindset (promotion focus). This occurs because consumers perceive AI as more artificial and lacking authenticity due to its mechanical nature, whereas they perceive humans as more authentic and grounded in genuine intent. Artificiality invites scrutiny for what might go wrong, heightening vigilance; authenticity invites engagement, fostering eagerness. This shift, in turn, carries into how consumers respond in the marketplace. Five experimental and archival studies support this account. Consumers implicitly associate AI with vigilance and humans with eagerness (Studies 1a–1b). In Studies 2a–2b, after browsing an AI- (vs. human)-based service website, consumers perform better at a prevention-related task (i.e., error detection) and worse at a promotion-related task (i.e., creative thinking), even though they rate AI and human providers as equally trustworthy, likable, and expert. In Study 3, consulting with an LLM-based business mentor (vs. a human) induces a prevention-focused mindset because AI seems more artificial and less authentic. As a consequence of this shift in regulatory mindset, Study 4 demonstrates a framing match: advertising for a retailer with AI-based customer service is less effective than for human-based customer service when framed in promotion terms (i.e., achieving gains), but this disadvantage disappears when the advertisement is framed in prevention terms (i.e., avoiding losses). Finally, an archival study (Study 5) shows that U.S. states populated by individuals demonstrating a proclivity toward a prevention- (vs. promotion)-focused mindset tend to exhibit higher levels of ChatGPT usage. Together, these studies suggest that practitioners who disclose AI involvement in their offering should incorporate prevention themes into their messaging, and that the cost of AI disclosure is likely greatest for offerings that sell aspiration and exploration.