Examining User Engagement in AI-based Interactional Tools: Valorizing the Essence of Human-Beings

Abstract

As the COVID-19 pandemic batters the world and disrupts the economic activities globally, companies are attempting to reshape the future by swiftly turning to use AI-based interactional tools to comply with the mandated social-distancing protocols. However, practitioners are unsure about the efficacy of AI-based interactional systems in facilitating the desired user engagement. Turing test premises user engagement in AI-tools based on the “indistinguishability” of AI-tools from human-beings. Such a user engagement approach for AI-tools is very different from the dominant user-engagement approach employed in prior IS research focusing on the “instrumental” value of technologies. Rooting our arguments in “human competency” and “media naturalness” literature, we theorize the significant role of the three human-like interactional competencies —cognitive, relational, and emotional competencies in interactional-AI tools, for fostering the desired user engagement with AI. We also hypothesize the mediating role of user trust in the relationship between ‘human-like interactional competencies in AI tools’ and ‘user engagement’. The theorized model is tested through a sequential mixed-methods approach comprising a quantitative two-wave survey followed by qualitative interviews with a sampled set of AI users. Results from our study largely support the theorized model. The emerging theoretical and practical implications are also discussed.