The accelerating integration of industrial robots into human workspaces under Industry 4.0 and emerging Industry 5.0 paradigms is reshaping the nature of decision-making on the factory floor. In human-cyber-physical systems (HCPS), where human cognitive flexibility is paired with robotic precision, the physical proximity and level of automation in human-robot collaboration have profound psychological implications. This study, involving 222 participants, examined how spatial distance and task-sharing with robots influence moral judgments in workplace scenarios.
Four distinct interaction levels, as defined by Bdiwi et al. (2017), framed the experimental design: Level 1 featured spatially separated tasks and zones; Level 2 introduced a shared task in a cooperation zone without physical contact; Level 3 involved shared workspace and direct handovers; Level 4 entailed shared workspace and joint tasks with direct physical contact. Participants encountered eight dilemmas—half life-or-death, half injury—adapted from the Footbridge dilemma but grounded in realistic industrial contexts. Responses were recorded on a four-point scale, distinguishing deontological decisions, driven by adherence to moral norms, from utilitarian decisions, focused on maximizing overall benefit.
The results revealed two strong patterns. First, life-or-death dilemmas elicited more deontological judgments, consistent with the emotional weight and cultural norms surrounding the sanctity of life. Injury dilemmas, by contrast, saw a higher incidence of utilitarian choices, suggesting a shift toward rational cost-benefit analysis in non-lethal contexts. Second, closer physical proximity between humans and robots correlated with increased utilitarian decision-making. This effect was pronounced across interaction levels, though levels 2 and 3 did not differ significantly, possibly due to their similar degrees of contact.
The dual-process theory of cognition offers a useful lens here. System 1, fast and intuitive, aligns with deontological reasoning, while System 2, slow and analytical, supports utilitarian reasoning. In high-proximity scenarios, participants may unconsciously adapt to the robot’s unemotional, rational operational style, engaging System 2 more readily. Overreliance on robotic partners and shifts in perceived responsibility—documented in earlier human-machine teaming research—could further diminish emotional involvement, nudging decisions toward utilitarian outcomes.
Interestingly, an unanticipated interaction emerged: in Level 3, life-or-death dilemmas prompted more deontological responses than injury dilemmas, diverging from the general proximity-utilitarian trend. This may reflect differences in how participants perceived the cause of the robot’s malfunction. When failures seemed random, as in some Level 1 and Level 3 scenarios, System 1 responses dominated. In scenarios where malfunctions were plausibly linked to the robot’s task motions, analytic reasoning was more accessible, fostering utilitarian choices.
The study’s strengths include its high statistical power and the introduction of injury dilemmas, which broaden the scope of moral judgment research beyond abstract, extreme cases. However, limitations must be acknowledged. The sample skewed young, educated, and predominantly female, raising questions about representativeness for industrial populations. The dilemmas were newly constructed, with internal consistency yet to be validated against established measures like the Footbridge and Trolley problems. Fixed scenario order and the artificiality of online surveys also constrain real-world applicability.
For engineers and designers of collaborative robots, these findings underscore the importance of considering psychological and ethical dimensions alongside technical safety measures. As physical separation diminishes and robots gain adaptive capabilities, human workers’ moral frameworks may shift in subtle but significant ways. Training programs could address potential overreliance and responsibility diffusion, reinforcing situational awareness and ethical accountability.
Future research could deepen these insights by manipulating perceived robot autonomy, incorporating richer media to simulate workplace contexts, and diversifying participant demographics. Exploring variables such as emotional intelligence or prior experience with automation may reveal further nuances in how humans navigate morally charged decisions in the presence of intelligent machines. As robots expand into domains like healthcare and service industries, understanding these dynamics will be critical to designing systems that support both performance and principled human judgment.