Impact of induced stress on individual performance, team coordination, and human-AI collaboration

Abstract

A range of applications involves human teams working under pressure on complicated tasks that require well-coordinated action planning to function effectively. This includes, for example, disaster response, where groups of emergency responders work together to locate casualties; air traffic control, where controllers need to predict potential accidents; and emergency operating rooms, where multiple clinicians operate on patients. Human performance is known to be affected by the level of stress that individuals are subjected to during the performance of tasks. The team members must have clear communication, mutual trust, and a shared understanding of the task to work efficiently and effectively under stress. It also needs to be clarified how stress affects individuals working in a team or the team as a whole. This thesis aims to uncover critical interactional and task-related elements that affect individual and team performance under stress. Specifically, we focus on stress induced by time pressure, performance pressure, and audio distraction and study how incorporating verbal and nonverbal cues impacts team coordination and performance. We conducted a series of experiments using online and in-person tasks. Firstly, a task was designed to monitor individual performance under time pressure and auditory distraction. The experiment was conducted remotely, measuring the performance of 32 participants. Our findings indicate that time pressure-induced stress positively impacts individual performance. Second, we divided participants into low, medium, and high-performing groups based on their overall performance in individual tasks into eight teams. Again, the experiment was conducted remotely using Google Sheets and measured the performance of all eight teams. The study found that time-pressure-induced stress negatively affects team performance. Additionally, we analyzed the coordination strategies used by high- and low-performing teams, showing that high-performing teams use more implicit coordination and have a high anticipation ratio. Third, we designed another experiment to understand the influence of automated agents on individual performance. This in-person experiment involved 32 participants competing against automated agents. The analysis of the results revealed that the automated agents' performance significantly influenced participants' performance. Participants' overall performance was slower when competing with slow agents, whereas competition with fast agents improved performance. Overall, our experiments provide insights into how stress and automated agents affect individual performance, team performance, and team coordination.

More information

Identifiers

ORCID for Sarvapali Ramchurn: orcid.org/0000-0001-9686-4302

ORCID for Sebastian Stein: orcid.org/0000-0003-2858-8857

Catalogue record

Export record