An Antisocial Social Robot: Using Negative Affect to Reinforce Cooperation in Human-Robot Interactions

Inspired by prior work with robots that physically display positive emotion (e.g., [1]), we were interested to see how people might interact with a robot capable of communicating cues of negative affect such as anger. Based in particular on [2], we have prototyped an anti-social, zoomorphic robot equipped with a spike mechanism to nonverbally communicate anger. The robot's embodiment involves a simple dome-like morphology with a ring of inflatable spikes wrapped around its circumference. Ultrasonic sensors engage the robot's antisocial cuing (e.g., “spiking” when a person comes too close). To evaluate people's perceptions of the robot and the impact of the spike mechanism on their behavior, we plan to deploy the robot in social settings where it would be inappropriate for a person to approach (e.g., in front of a door with a “do not disturb” sign). We expect that exploration of robot antisociality, in addition to prosociality, will help inform the design of more socially complex human-robot interactions.