Can a robot invigilator prevent cheating?

One of the open questions in Educational robots is the role a robot should take in the classroom. The current focus in this area is on employing robots as a tool or in an assistive capacity such as the invigilator of an exam. With robots becoming commonplace in the classroom, inquiries will be raised regarding not only their suitability but also their ability to influence and control the morality and behaviour of the students via their presence. Therefore, as a means to test this cross-section of Educational robots with the underlying issue of morality and ethics we conducted an empirical study where the Nao robot invigilated an exam for a group of students. A between-subjects design (N = 56, 14 groups of 4 students) compared whether Nao was able to deter students from cheating and maintaining their discipline in comparison to a human invigilator or when there was no invigilator present. Our results showed that while explicit cheating rarely took place across all conditions, the students were significantly more talkative when they were invigilated by Nao. In conclusion, we discuss and speculate upon some of the ensuing implications towards not only the application of robots in education but also consequently the wider issue of the preservation of morality and ethics in a classroom in the presence of an agent.