As virtual world games become popular avenues for emotional relief, artificial intelligence (AI)-powered non-player characters (AI-NPCs) are redefining gaming experience with their unparalleled intelligence and adaptability different from traditional NPCs. However, how AI-NPCs affect players' subjective well-being remains underexplored. This paper addresses this gap by elucidating how AI-NPCs affect players' subjective well-being, with Antecedent-Belief-Consequence framework as a guiding lens.
This paper adopts a sequential mixed-method research design. In the qualitative phase, the affordances of AI-NPCs relevant to players' subjective well-being were identified through 20 semi-structured interviews. Built upon the identified affordances, in the quantitative phase, a theoretical model was built and a total of 326 valid responses were obtained in the online survey.
The results have identified four technology affordances, namely interactivity affordance, emotional support affordance, aesthetic affordance and narrative affordance. Among them, interactivity and emotional support affordances are prominent antecedents of social identification, while aesthetic and narrative affordances enhance game identification. Both social identification and game identification contribute to players' subjective well-being.
Lying at the intersection of technology affordances and social identifications, our work is the first to investigate the impact of AI-NPCs design on players' subjective well-being in virtual world games. It not only extends the theory of affordances and identifications into the virtual world but also provides practical implications to facilitate effective human–AI interactions.
