Who Makes the Best Robot Teammate? New Study Reveals Surprising Gender, Age and Skill Gaps

When a human and a robot share a workbench, the biggest bottleneck is not the machine. It is the human brain, forced to juggle competing demands at once. A new empirical study from Ben-Gurion University of the Negev, published in Frontiers in Neurorobotics, has put that juggling act under the microscope in one of the first investigations to combine dual-task interference with industrial human-robot collaboration. The findings challenge long-held assumptions about expertise, expose an unexpected gender gap in teamwork, and suggest that the fastest worker on the floor may be the worst robot teammate.

Dual-task interference is a well-documented limitation of human cognition: when the brain must rapidly switch attention between multiple tasks attempted simultaneously, performance on one or both tasks degrades. It is the same phenomenon that makes smartphone notifications so costly, or that makes talking on a phone while driving dangerous. Yet despite decades of research in psychology and human-computer interaction, the authors report that no prior study had examined how this interference plays out when a human must coordinate with collaborative robots in a time-bound industrial assembly setting, precisely where stochastic interruptions and competing cognitive demands are everyday realities.

To fill that gap, the research team designed a cognitive joint-assembly task involving 172 participants. Each person collaborated with two identical five-degrees-of-freedom lightweight robot manipulators equipped with parallel grippers, while simultaneously solving a logic-based Gravity Maze puzzle that required visual perception and problem-solving. The robots operated on independent cycles of 30 and 60 seconds, entering the shared workspace to collect finished Lego assembly pieces from the human. Because the intervention timing depended on the interplay of robot waiting times, human waiting times and the increasing complexity of the maze, the interference load was unpredictable, peaking during double interventions when both robots demanded assistance at once.

The experimental design was deliberately unforgiving. Participants were told that the clock would run twice as fast whenever the left robot was waiting for them, three times as fast whenever the right robot was waiting, and six times as fast when both robots were waiting simultaneously. These time-acceleration penalties emulated the real cost of robot idle time on team productivity and forced subjects to distribute their attention across the primary cognitive task and the secondary robotic handovers. The coordination protocol was simple: the human never waits, the robot may need to wait. Success therefore depended entirely on the collaboration skills of the human partner.

Performance was measured with three objective metrics: the number of human-robot interactions, which gauged user engagement; total robot idle time, which gauged team fluency; and the number of completed cognitive tasks, which gauged primary task performance. The population variables were gender, age group and proficiency, with adolescents aged 13 to 14 compared against university students aged 21 to 29, and experts with prior job-relevant experience compared against novices. Crucially, the experts were proficient in the individual tasks but had no prior experience with the combined dual-task environment, ensuring the interference effect remained a novel challenge for everyone.

The gender results were striking. Female participants exhibited 12 percent lower robot idle times and 14 percent more human-robot interactions than males, indicating greater collaboration with the robots and more fluent teamwork. Yet there was no significant difference between the genders in the number of cognitive tasks completed. The authors interpret this as evidence that females distributed their attention equivalently across the primary and secondary tasks, valued the cost of a waiting team partner more highly, and worked more collaboratively toward the overall team objective. This finding contradicts earlier social robotics research suggesting men typically show higher engagement in cooperative close-proximity tasks, and it aligns instead with industrial manufacturing studies reporting that female subjects were more productive at assembly workstations, reinforcing calls for gender-adjusted production line models.

The age comparison produced an equally counterintuitive twist. Adolescents completed their primary cognitive tasks significantly faster than adults, outperforming them by 21 percent and even beating the experts by 22 percent in the number of successful assemblies. Their superior problem-solving efficiency and fact-recall processes translated into a higher learning rate and better management of the dual-task demands. However, this raw speed came at a hidden price: adults and experts demonstrated significantly greater team fluency, with roughly 12 percent lower robot idle times. The researchers suggest that the adolescents’ intense focus on primary task momentum meant each robot interaction broke their rhythm, imposing the temporal and cognitive cost of task switching and producing higher variability in handovers. In other words, higher performance speed in the primary task negatively affected the fluency of the human-robot team, an inherent trade-off between task throughput and the cognitive costs of coordinating concurrent work.

Perhaps the most humbling result concerned expertise. No significant differences were observed between novices and experts on any of the three performance measures. The dual-task environment acted as a kind of equalizer, erasing the cognitive edge that primary task proficiency and robot familiarity were expected to confer. The authors found that novices successfully bridged the skill gap by adopting a more balanced attention distribution strategy, allocating equivalent focus to the maze and the robot interventions and thereby minimizing the cumulative time penalties. This suggests that in dual-task collaboration, strategic coordination rather than solo task proficiency is the primary driver of success. The study concludes that task proficiency and team coordination are two distinct and mutually independent skills, yielding a four-tier user model: expert or novice in the primary task, crossed with good or poor team player. A good team player, the authors note, can offset the edge of a well-acquainted expert.

Correlational analyses reinforced these conclusions. The number of interactions correlated strongly and negatively with robot idle time, showing that subjects who valued a waiting partner’s cost were more collaborative and more successful in the team objective, with the female cohort as the clearest example. The number of completed cards also correlated negatively with idle time, because every second the robot waited accelerated the clock and shrank the time available for the primary task. Together, the results reveal that effective team coordination skills are as important as task proficiency in dual-task human-robot collaboration.

For the robotics industry, the implications are concrete. The authors demonstrate that collaborative robot learning algorithms can classify users into specific population clusters based on their characteristic patterns of interaction, allowing systems to adapt control strategies in real time for improved performance and user experience. They recommend that designers monitor primary task speeds and use visual or auditory notifications to alert workers when their pace threatens team fluency, facilitating human-robot mutual adaptation. They also argue for workforce training modules targeting dual-task coordination skills specifically, since primary-task training alone proved largely irrelevant in this environment, and since dual-task skills can be acquired, retained and transferred across tasks. As small and medium enterprises and less automated factories such as semiconductor and aerospace plants introduce collaborative robots for tasks previously done solely by humans, understanding who makes the best robot teammate, and why, may become as important to productivity as the robots themselves.

Subject of Research: Designing for gender, age & proficiency: an empirical study of the cognitive ergonomics of dual-task interference in industrial human-robot collaboration

Article Title: Designing for gender, age & proficiency: an empirical study of the cognitive ergonomics of dual-task interference in industrial human-robot collaboration

Article References: Roy, S., Peretz, I., Sayfeld, L., & Edan, Y. (2026). Designing for gender, age & proficiency: an empirical study of the cognitive ergonomics of dual-task interference in industrial human-robot collaboration. Frontiers in Neurorobotics, 20, Article 1222566. https://doi.org/10.3389/fnbot.2026.1222566

Image Credits: AI Generated

DOI: 10.3389/fnbot.2026.1222566

Keywords: Designing, gender, proficiency, empirical, cognitive, ergonomics, dual-task, interference, industrial, human-robot, collaboration, scientific research