The ERROR-MM Project: Exploring Robotlike Robot Behaviours in Users’ Mental Models

Sandra MC Andrade, Margarida Fortes-Ferreira, Friederike Eyssel, Nikolas Martelaro, Isabel Neto, Ivo Roupa, Ricarda Wullenkord, Filipa Correia

Keywords: Social Robotics, Human-Computer Interaction, Robot Behaviour, Robomorphism

TL;DR: The ERROR-MM project will explore robomorphism and test its effect on the accuracy of human mental models of robots.

Abstract: Design aims to align the user's mental model with the actual behaviour of a device to reduce the gap between user expectations and observed behaviour. In robotics, because of the anthropomorphic attributions that humans make towards robotic machines, humanlikeness has been one of the main heuristics to guide the design of robots, regarding their embodiment as well as their behaviours. Although humanlike design has many proven advantages, it also comes with disadvantages, such as being potentially deceitful regarding the robot's true capabilities. Recently, a new paradigm of robomorphism has been proposed that speculates on the attribution of robotlikeness to things. The ERROR-MM project will explore this novel idea of robotlikeness and test its effect on the accuracy of human mental models of robots. Specifically, starting with movement, as it is the basis of any robotic behaviour, the project aims to test the hypothesis that a robot employing robotlike behaviours will lead humans to form better mental models of a robot's true capabilities than a robot that employs humanlike behaviours.

Submission Number: 8