Human-Artificial Intelligence Collaborative Decision-Making in Emergencies: Relative Advantage Theory

ObjectiveThis study proposes Relative Advantage Theory to explain decision-making mechanisms in emergencies where humans collaborate with AI systems, and provides initial empirical evidence consistent with its core assumptions.BackgroundWhile artificial intelligence (AI) can enhance decision-making efficiency and accuracy in emergencies, the mechanisms underlying human use of AI in such contexts remain poorly understood.MethodParticipants completed a simulated emergency response task. The experiment employed a 2 (time pressure: with vs. without) × 2 (AI role: expert vs. assistant) within-subjects design, measuring perceived self-capability, perceived AI capability, compliance, and electroencephalography (EEG) activity during the task.ResultsTime pressure and AI role influenced compliance through relative advantage (perceived AI capability minus perceived self-capability). Higher relative advantage values predicted higher compliance rates.ConclusionFindings across different time pressures and AI role support Relative Advantage Theory, which provides a potential explanation for inconsistencies reported in prior studies.ApplicationThe findings offer theoretical insights for designing AI systems optimized for emergency use.