Presented at the IEEE/ACM International Conference on Human-Robot Interaction, this study by Meiying Qin, B. Scassellati, and Laurie R. Santos explored canine social cognition in an emerging domain: interactions between dogs and robots. Pointing comprehension is a fundamental measure of social understanding in dogs, yet little work has examined how dogs respond when the pointer is an artificial agent rather than a human.
Earlier research showed that dogs did not follow pointing gestures from a non-humanoid robot. To assess whether human-like morphology might influence dogs’ responses, the authors tested 20 domestic dogs using the humanoid robot Nao. The robot produced standard pointing gestures while dogs were asked to choose between two potential food locations.
The findings revealed no reliable evidence that dogs understood or responded appropriately to humanoid robot pointing. Unlike human pointing—which dogs typically follow with high accuracy—robotic gestures did not elicit consistent directional choices. This suggests that dogs may not readily attribute agency or communicative intent to robots, even when the robot approximates human form.
The researchers emphasize that this is an early stage of inquiry. They are currently refining the experimental design and plan to conduct further studies to explore how factors such as movement dynamics, sound, familiarity, or hybrid human–robot interactions might influence dogs’ perception of robotic agents.
This work contributes to growing interdisciplinary interest in animal–robot interaction, with implications for future applications in training, enrichment, and automated care systems for companion animals.
Source: Qin, M., Scassellati, B., & Santos, L. R. (2019). Agency in Canine-Robot Interaction: Do Dogs Understand Humanoid Robots’ Pointing Behavior? Presented March 1, 2019.
