Falling out with AI-buddies: The hidden costs of treating AI as a partner versus servant during service failure

Abstract

The swift integration of artificial intelligence (AI)-driven tools in various industries, such as virtual assistants, chatbots, and service robots, raises inquiries about consumer reactions to these emerging technologies. To promote acceptance and enhance service interactions, companies frequently market these technologies by fostering parasocial and anthropomorphic relationships: the roles of partner and servant are among the most prevalent. Yet, the precise influence these relationship roles have on consumer responses remains uncertain. While extant literature primarily shows a positive effect of treating AI as a partner, in the current research, we find a multifaceted adverse effect of anthropomorphic partner (versus servant) relationships in the context of service failure. Across four studies, the results demonstrate that when consumers perceive an AI assistant as a relational partner, it heightens their inclination to attribute the failure to themselves because of elevated self-expansion perceptions with the AI. Furthermore, within this relationship dynamic, users exhibit reduced intentions of utilizing the AI agent again, as a result of a decreased sense of self-efficacy. Finally, the undesirable effects of a partner relationship following a service failure can be mitigated by drawing attention to the AI's learning capabilites. The findings of our research highlight a potential caveat of an AI-as-partner relationship, thus advancing our understanding of consumer interaction with AI from a relational perspective. • Framing AI as a partner (vs. servant) increases consumer self-blame after service failure. • This is due to greater self-expansion when the AI is positioned as a partner. • Partner-AI framing lowers reuse intention after failure due to reduced self-efficacy. • Highlighting AI’s learning ability reduces negative reactions to partner-AI failure.