Shaping Empathy and Trust Toward Agents: The Role of Agent Behavior Modification and Attitude

As artificial agents become increasingly integrated into daily life, designing interactions that foster empathy and trust is vital for ensuring ethical and effective human-agent collaboration. This study investigates how an agent’s attitude (positive or negative) and behavior modification (positive, negative or none) jointly influence users’ perceptions of empathy and trust over time. Using a traffic safety reflection scenario, we conducted a three-factor mixed design experiment involving 582 participants, who interacted with a computer-animated agent and evaluated it both before and after behavioral changes. The results revealed that positive behavior modification significantly enhanced empathy and trust, particularly when the agent’s attitude remained positive throughout. Interaction effects between behavior modification and time (before/after) were statistically significant (p <.001), and agents with a consistently positive attitude were rated significantly higher on both empathy (M = 3.65) and trust (M = 41.19) than negative ones. These findings underscore the importance of emotional responsiveness and adaptive behavior in shaping social perceptions of agents. The study contributes to theoretical frameworks of human-agent trust and empathy by highlighting how temporal dynamics and affective cues influence relational outcomes. These insights offer practical guidance for designing agents that are perceived not only as intelligent but also as emotionally attuned and socially trustworthy.