Effects of Human–Automation Authority Allocation on Multitasking Performance under Workload Conditions

The increasing autonomy of intelligent cockpit systems raises critical questions about how authority should be allocated between humans and machines, particularly under dynamic workload conditions. This study employed the multi-attribute task battery to examine two authority allocation strategies: human-led authority allocation (HLAA) and shared authority allocation (ShAA). In Experiment 1 (N = 23), we validated workload manipulations and confirmed that continuous tracking tasks were most sensitive to workload increases. In Experiment 2 (N = 69), we compared HLAA and ShAA with advanced automation participation under fluctuating workload transitions. Results revealed a direction-specific interaction: ShAA supported better performance during low-to-high workload transitions, while HLAA was more effective during high-to-low transitions and yielded lower subjective fatigue. Although ShAA leveraged automation to stabilize escalating demands, it also introduced conflict and delayed human re-engagement. These findings highlight the need for adaptive, context-sensitive authority allocation mechanisms and provide design guidelines for enhancing transparency, trust calibration, and workload management in next-generation single-pilot cockpits.