Robot Friendship: Can a Robot be a Friend?

Introduction

The question posed here is one about an aspect of human cognition and emotion that is often overlooked or ignored, namely friendship, and its possible instantiation in artificial systems such as robots. Can robots be friends? Probably for most people it seems a bit weird to suggest so, or to imagine that humans and robots can enter into interactive dynamics similar to friendship. The intuition that something would be missing for two artificial systems really to be ‘friends’ may well be right, but as always with intuitions in the context of research they need to be articulated, and articulating this intuition – or counter-claims that robots could not be friends – demands a deeper understanding of not only the specificities of human cognition, but also the relation between cognition and affection, and the nature of friendship. This article focuses on the later.

The claim of this paper is not that friendships between robots, or between humans and robots, are impossible. The suggestion is to keep the question – about possible future realizations of friendship in human-robot or robot-robot relations – open, and investigate more in detail what friendship among humans might tell us about similarities and differences between natural human cognition (including distributed cognition) and artificial robotic cognition. Friendship and social cognition in general has been seen as important for the development of an autonomous self in humans, and this may thus teach us something about human autonomy and agency. In the fields of Cognitive Science, Robotics, AI and Artificial Life, the focus has often been one of building adaptive systems intelligent enough to solve practical tasks such as survival, transforming sensory information into organized knowledge to guide movement, and learning instrumental skills needed for supporting human action and cognition. This kind of instrumental rationality and intelligence is important for individual actions in many technical domains, but may not be sufficient for all social actions in domains where values and norms play an important role. By asking questions about forms of human friendship and the possible existence or non-existence of robot friendship, we also ask what it means to say that rational purposeful human action in the social domain is guided by norms. Such norms are of a social character, they are expressions of self-control, and they regulate not just the action of isolated individuals, but emphatically social individuals, that is, individuals socialized through their development, like social organisms of other species, or like the ‘political animals’ (Aristotle) we are qua living in societies with social institutions.