Recent advancements in technology have enabled robots to work alongside humans. While this has introduced a new dynamic of teamwork, researchers from the Technical University of Berlin have discovered that humans might be prone to “social loafing” when working with robots.
Social loafing is a phenomenon where individuals put in less effort when working in a group, especially if they believe their contributions will not be noticed or are overshadowed by a high-performing team member. The study, led by Dietlind Helene Cymek and published in Frontiers in Robotics and AI, aimed to discern whether humans displayed this behavior when paired with robots.
The team set up a defect-inspection task where participants had to identify errors on circuit boards. The boards were intentionally blurred, and participants could only view clear images by maneuvering a mouse tool over them. This allowed the research team to monitor each participant’s inspection process.
Half of the participants believed they were inspecting boards previously checked by a robot named Panda. Although they never worked directly with Panda, they were aware of its presence and could hear it in the background. Post-experiment, all participants were required to evaluate their effort, sense of responsibility, and overall performance.
Upon initial assessment, Panda’s presence seemed inconsequential. Both groups spent similar time inspecting the boards and rated their sense of responsibility, effort, and performance alike. However, a deeper look revealed that the group “working with” Panda exhibited a decline in defect detection as the task progressed. This suggests that they may have subconsciously relied on Panda’s accuracy, leading to a “looking but not seeing” effect.
Dr. Linda Onnasch, the study’s senior author, remarked, “While we can easily track where someone is looking, determining if they’re mentally processing the visual information is a challenge.”
The study highlighted potential safety concerns, especially in safety-critical domains where double-checking is paramount. A decrease in motivation and attention, even over a short 90-minute task, can have significant ramifications for work outcomes in longer shifts.
However, the researchers acknowledged the study’s limitations. Participants were aware they were being observed, which might affect genuine social loafing. Furthermore, they never directly interacted with Panda. As Cymek noted, “To truly gauge the extent of motivation loss in human-robot interaction, we must test our hypotheses in real-world settings with professionals accustomed to working with robots.”
