Enhancing Perceived Empathy in Empathic Mixed Reality Agents via Context-Aware Adaptation

Area of Science:

• Human-Computer Interaction
• Virtual Reality
• Empathic AI

Background:

• Mixed Reality Agents (MiRAs) are key for virtual-physical interaction.
• Perceived empathy in MiRAs is under-researched, limiting applications in therapy and education.
• Empathic Mixed Reality agents (EMiRAs) offer potential for enhanced human-agent interaction.

Purpose of the Study:

• To investigate the impact of an Empathic Mixed Reality agent (EMiRA) on user social perceptions.
• To analyze how EMiRAs adapting to physiological and physical states affect user experience.
• To explore design implications for future empathic agents in mixed reality.

Main Methods:

• Developed an EMiRA that dynamically adapted to user physiological data and in-game physical events.
• Evaluated user perceptions of the agent within a shooting game context.
• Measured social presence, social connectedness, perceived empathy, paternalism, and user dominance.

Main Results:

• EMiRA adaptation significantly enhanced users' social presence, social connectedness, and perceived empathy.
• Physiological adaptation led to increased perceptions of paternalism and reduced user dominance.
• Adaptation to physical events did not yield significant changes in user dominance or paternalism.

Conclusions:

• Adaptive EMiRAs can improve social perceptions and user experience in mixed reality.
• Physiological adaptation influences social dynamics, suggesting careful design considerations.
• Findings provide valuable insights for developing more effective and empathetic AI agents.