Scenario, Role, and Persona: A Scoping Review of Design Strategies for Socially Intelligent AI Agents
Ruowen Niu, Jiaxiong Hu, Siyu Peng, Caleb Chen Cao, Chengzhong Liu, Sirui Han, Yike Guo
Abstract
As artificial intelligence (AI) agents have often been perceived as social actors, there is a growing expectation for them to demonstrate social intelligence. Social intelligence encompasses the ability of AI agents to align their actions with human intentions and behave in socially and culturally appropriate ways. This study conducts a scoping review to analyze the design strategies employed in current AI agent designs that support the exhibition of social intelligence. Our findings reveal three design strategy themes that interdependently and collectively shape the capabilities and behaviors of AI agents: scenario, role, and persona. These findings provide a structured perspective and actionable insights for designing AI agents that effectively integrate social intelligence into human-AI interactions.