Security robots are increasingly appearing in public places: they patrol parking lots, keep order and can respond to requests for help. As their functions expand, a question arises: are people ready to perceive a machine as a boss? Researchers from the University of Michigan School of Information - doctoral student Xin Ye and Professor Lionel Robert - sought an answer.
The experiment involved 100 people. The scientists tested how two types of justice affect trust. The first is interpersonal justice: whether the robot gives a person the opportunity to explain and explains its decisions. The second is distributive justice: whether the robot applies the rules equally to different groups. The results were published in the journal Proceedings of the Human Factors and Ergonomics Society Annual Meeting.
It turned out that the robot's manner of communication and consistency in applying the rules increased participants' trust. Fair interaction also increased willingness to recognize the robot as an authoritative figure. At the same time, this is a statistical association: the results do not imply that justice directly causes trust - it may be associated with it alongside other factors.
Xin Ye notes: in human-robot interaction research, machines are more often studied as assistants or partners. She is interested in a different scenario - when a robot receives power over people. The work was based on research into justice in policing: the scientists tested whether people transfer similar expectations to robot bosses.
The experiment took place in a parking lot security scenario. According to the authors, in situations with higher stakes people's reactions may differ. Nevertheless, the data obtained provide a starting point for designing such systems. "If we are going to develop and deploy robots in these roles, we need to understand that fairness matters," says Xin Ye.
The researchers recall that security robots are already used in practice. For example, the New York Police Department used a robot from Knightscope, though it was later taken out of service. Such cases raise questions about how such systems should be designed and how people react to them.