As technology becomes increasingly integrated into daily life, understanding how both humans and animals perceive artificial agents is vital for designing socially intelligent systems. In this exploratory study presented at the 8th International Conference on Human–Agent Interaction, Haruka Kasuga and Yuichiro Ikeda investigated how dog owners and their dogs responded to different robotic agents, including the dog-shaped smart speaker, the humanoid robots NAO and Pepper, and the voice assistant Google Home.
The research involved 33 dog owners and 21 dogs in two experiments. The first evaluated human impressions of the agents, while the second analyzed dogs’ behavioral responses when exposed to the same devices. Using the GODSPEED questionnaire, owners rated the robots on dimensions such as anthropomorphism, animacy, likability, intelligence, and safety. Results showed that owners perceived humanoid robots as significantly more intelligent and appealing than the dog-like smart speaker.
However, dogs reacted quite differently. Canine participants showed more social behaviors—such as sniffing or approaching—toward the dog-like speaker than toward humanoid robots or standard smart speakers. This suggests that while humans assess artificial agents based on perceived sophistication, dogs respond to more familiar, biologically relevant cues.
The findings underscore a key insight: the social meaning of technology differs across species. For humans, interaction with humanoid robots may evoke cognitive empathy and novelty, whereas for dogs, familiar shapes and sounds may elicit genuine social interest. Kasuga and Ikeda highlight that this divergence provides an essential foundation for designing multi-species human–animal–technology interactions that respect both human expectations and animal perception.
Source: Kasuga, H., & Ikeda, Y. (2020). Gap between Owner’s Perceptions and Dog’s Behaviors Toward the Same Physical Agents: Using a Dog-like Speaker and a Humanoid Robot. Proceedings of the 8th International Conference on Human–Agent Interaction. Published November 2, 2020.
