Impacts of physical and informational failures on worker-autonomy trust in future construction

Autonomous agents are increasingly becoming construction workers' teammates, making them an integral part of tomorrow's construction industry. Although many expect that worker-autonomy teaming will enhance construction efficiency, the presence of auto-agents, or robots necessitates an appropriate level of trust-building between workers and their autonomous counterparts, especially because these auto-agents' perfection still cannot be guaranteed. Although researchers have widely explored human-autonomy trust in various domains - such as manufacturing and the military - discussion of this teaming dynamic within the construction sector is still nascent. To address this gap, this paper simulated a futuristic bricklaying task to (1) examine whether identifying autonomous agents' physical and informational failures and risk perception affect workers' trust levels, and (2) investigate workers' neuropsychophysiological responses as a measure of trust levels toward robots, especially when autonomous agents are faulty. Results indicate that (1) identification of both types of failures and high-risk perception significantly reduce workers' trust in autonomous agents, and the nuances of workers' responses to both types of failures were discerned; and (2) brain activation correlates with trust changes. The findings suggest that workers' unfamiliarity with autonomous technologies, coupled with fast-growing interest in adopting them, may leave workers at risk of improper trust transfer or overtrust in the autonomous agents. This study contributes to an expanding exploration of worker-autonomy trust in construction and calls for further investigations into effective approaches for auto-agents to communicate their physical and informational failures and to help workers recover and repair trust.