Human preferences for cognitive and emotional capabilities in robots across different application domains

People's preferences regarding cognitive and emotional capabilities in robots need to be considered in the design of robotic systems that align with human values and expectations. This study investigates how such preferences vary across different robotics application domains and identifies key influencing factors. In a between-subjects study with 271 participants, both quantitative and qualitative data were collected on preferences for 12 mental (cognitive and emotional) capabilities in six types of robots, each situated in a specific domain: healthcare, defense, household, social, education, or customer service. The results reveal a general preference for agency-related abilities (e.g., planning, reasoning) over experience-related abilities (e.g., feeling happiness, pain) across all domains. However, there was a weaker preference for agency capability in household cleaning robots and a stronger preference for experience capability in social companionship robots. Qualitative analysis revealed a common desire for robots to function objectively and logically, without emotions, while still showing empathy toward human mental states. Additionally, gender and educational background emerged as factors influencing participants' preferences. Unlike previous research, which mainly focused on the attribution of mental capabilities to robots, this study offers insights into human preferences and the factors shaping them, which can inform the design of future robots and help facilitate their successful integration into society.

Hilda Nääs, Sam Thellman, Tom Ziemke