Customer reactions to generative AI vs. real images in high-involvement and hedonic services

Given the emerging opportunities of generative AI for business and marketing, many companies are wondering whether they should use images created through generative AI for commercial purposes. Prior research on hospitality communication has not solved this issue, as AI-generated images are occasionally promoted as effective marketing tools across various service contexts, while other scholars caution against their use due to the significant concerns they may trigger among consumers. Following a mixed-methods approach to find boundary conditions, our research reveals that consumers prefer hospitality services advertised with real images, rather than those featuring AI-generated images. Nevertheless, this effect is moderated by two key factors. In particular, the research reveals that the negative influence of using generative AI images on intentions to use and recommend the service are strengthened for hedonic rather than utilitarian services, and for highly rather than lowly involved customers. A qualitative study further explores the reasons behind this rejection, highlighting that customers perceive companies using AI-generated images as impersonal, less professional, lacking credibility, and potentially misleading, as they impede customers’ ability to envision the actual experience. Implications for management suggest that, while generative AI holds promise for enhancing communication, companies should use AI-generated images with caution. The discussion also proposes future research directions to explore the broader implications of AI use in marketing. • Customers avoid services advertised by AI-generated images versus real images. • The rejection of AI is greater for hedonic services and high-involvement decisions. • Customers distrust AI-generated images, which are perceived as lacking veracity. • These results are examined through mixed-methods, quantitative, and qualitative. • Companies should use real or AI-generated images depending on the researched factors.