The ethics of driverless cars

Abstract

This paper critiques the idea of full autonomy, as illustrated by Oxford University's Robotcar. A fully autonomous driverless car relies on no external inputs, including GPS and solely learns from its environment using learning algorithms. These cars decide when they drive, learn from human drivers and bid for insurance in real time. Full autonomy is pitched as a good end in itself, fixing human inadequacies and creating safety and certainty by the elimination of human involvement. Using the ACTIVE ethics framework, an ethical response to the fully autonomous driverless cars is developed by addressing autonomy, community, transparency, identity, value and empathy. I suggest that the pursuit of full autonomy does not recognise the essential importance of interdependencies between humans and machines. The removal of human involvement should require the driverless car to be more connected with its environment, drawing all the information it can from infrastructure, internet and other road users. This requires a systemic view, which addresses systems and relationships, which recognises the place of driverless cars in a connected system, which is open to the study of complex relationships, both networked and hierarchical.