Amani Alabed

Project description

Despite advancements in artificially intelligent (AI) agents, scant research examines the extent to which interactions with such humanlike agents resemble interpersonal relationships. In my research, I studied how consumers relate to anthropomorphised AI, specifically conversational AI, via their self-concept.

This exciting topic proposes several opportunities, yet challenges in unpacking the impact of consumer-AI relationships on one’s existing relationships with other humans. Toward this goal, I propose a conceptual model grounded in prior literature on psychology, human-robot interaction, and consumer behavior. The model explains how anthropomorphism contributes to perceiving AI agents as similar to oneself, i.e., self-congruence. Building on the theories of self-expansion and self-extension, the thesis introduces a novel concept, self-AI integration. This concept encapsulates the process in which users integrate the characteristics of AI agents into their self-concept. The model’s main relationships were tested using a mixed-method approach, which utilised qualitative and quantitative studies. The results of the studies reveal how AI anthropomorphism helps users position these agents in social roles like friends, personal secretaries, or even one’s mirror, which in turn impacts the way they form relationships with AI that to some extent replace human-to-human relationships. The experimental phase further revealed fresh insights regarding the role of novelty in shaping consumer-AI relationships.

This research presents insightful implications for researchers and practitioners to stand out in today’s dynamic marketplace. It draws attention to the psychological processes of self-congruence and self-AI as key concepts that explain how users can meaningfully relate to such AI agents via their self-concept and build relationships with them. It also guides managers on the type of promises to communicate with users to promote meaningful consumption experiences and build stronger consumer-AI bonds. 

This research was presented at two international conferences (EMAC and AIRSI), and its main conceptual model was published in the Technological Forecasting and Social Change Journal. The empirical studies are now being prepared to be published in academic journals.