Mental models of robotic assistants

Abstract

If robotic assistants are to be successful, people will need appropriate mental models of what these robots can do and how they operate. We are developing techniques for measuring people's mental models of interactive robots and social agents. We aim to measure the content of these models--if they are anthropomorphic or mechanistic--and the richness of these models (how elaborate or sparse they are; how much confidence people have in them). We report progress here.