A study by researchers at SWPS University found that elementary school children responded politely to a humanoid robot regardless of whether it addressed them politely or in a commanding manner. According to the authors, this suggests that established social norms influenced children’s behaviour more strongly than the robot’s communication style. The study also showed that younger children and girls were more likely to attribute human-like qualities to the robot, particularly when it used a polite manner of speaking.
The research, published in Computers & Education, examined how 251 children aged 7–12 interacted with Pepper, a 120-centimetre humanoid robot equipped with sensors, cameras and microphones. The robot was programmed to communicate either politely or in a commanding tone and was assigned a male or female name to test whether these variables influenced children’s responses.
During the experiment, children observed the robot imitating animals and reacted to its request, or refusal, to be photographed. They then answered questions about Pepper’s abilities, including whether it could experience emotions or imagine things.
The researchers reported that polite robots were more frequently perceived as having human-like qualities than commanding robots. This tendency increased when the robot’s politeness aligned with stereotypical expectations related to its assigned gender. The highest level of anthropomorphisation occurred when the robot was presented as female and polite.
The authors argue that communication cues shape how children perceive and engage with robots in educational settings. They note that adjusting robots’ interaction styles to children’s developmental levels and social expectations may support learning engagement.
The findings were presented during the HumanTech Summit 2025, held at SWPS University in Warsaw. The conference gathered researchers focusing on human-robot interaction and the social and psychological aspects of new technologies. Speakers included Emily Cross from ETH Zürich, Jessica M. Szczuka from the University of Duisburg-Essen, and Peng Liu from Zhejiang University.
Photo credit : SWPS University
