Safety and social credibility

An assistive robot is typically required to perform both safety and social functions, including interacting with a end-user and alerting them to hazardous situations. In order to gain end-user acceptance the robot must meet some minimum threshold of social credibility sufficient to encourage the end-user to interact and engage with it. Similarly, a certain degree of social credibility will be required for the end-user to correctly respond to an assistive robot's safety alerts.
In some situations the safety and social requirements for an assistive robot are likely to conflict, particularly where satisfying a safety requirement means the robot must act in a socially-inappropriate fashion. In this project we identify and characterise the link between social credibility and safety. We propose the foundations for a methodology for assuring safety of autonomous systems in the presence of required social characteristics.