Ripeng Feng Undergraduate Dissertation 2014/15

Can a robot understand?

Supervised by A.Sharkey

Abstract

There are an increasing number of robots currently being designed and produced, which can interact with people and fulfil special goals or objectives. This paper describes an experimental study designed to explore whether it is possible for human beings to be, in effect, tricked or duped by an animated robot. The experiment study asks people who have interacted with animated robot to rate the attributes of these robots, in accordance with a number of key queries (for example, 'how life like do you think the robot is?'). There are two parts to the experiment study. The first part asks participants to interact with the animated robot and to rate the attributes of the animated robot while the animated robot was activated. The second part asks the participants to complete the same tasks but this time, with the deactivated animated robot. The underlying presumption is that there will have a confusion made by the animated robot which will affect participants have a tendency to rate the animated robot more positive life like attributes. This will be offended by their answers.