Effects of proactivity, autonomy, and transparency on user perception in human-AI agent interactions

Abstract

As large language model-driven AI agents evolve from generative tools into collaborative entities with autonomous capabilities, the human-AI interaction paradigm is shifting from command execution to task delegation. However, empirical evidence on how the behavioral characteristics of AI agents reshape user perception remains limited. This study employs a within-subject experiment to examine the effects of proactivity, autonomy, and transparency on perceived usefulness, trust, expectation confirmation, satisfaction and continuance usage intention in a travel-planning task. The results show significant effects of proactivity and transparency across all five subjective outcomes, whereas autonomy shows no significant main effect. AI agents with high proactivity and low autonomy received the most favorable subjective evaluations; comparatively, AI agents with high proactivity and high autonomy were associated with lower continuance usage intention. Transparency plays an important role in maintaining user trust across conditions. This study offers preliminary and context-specific design considerations for balancing initiative, decision authority, and transparency in human-AI agent interactions.