Automated vehicle driving: background and deduction of governance needs

Abstract

Autonomous vehicles (AV) aim to absolve humans from cumbersome performance of tasks and eliminate them from being of a hazard source. Some state that AV will become reality soon, diffuse faster than classical automotive innovations, others are even afraid of hype. Platoons on highways and public transport vehicles coping with limited driving situations at low speed will lead the way, rather than general-purpose fully-automated cars; a combination of driverless and human-driven cars over decades is foreseeable. AV are expected to provide various benefits for users and our society. Although significantly improved traffic safety is expected severe crashes will happen, rejecting the vision of “zero risk”. Benefits are confronted by some “buts”: Potentially new stress due to blind reliance on algorithms and loss of control; wireless communication channels between vehicles and infrastructure may introduce systemic failures and new entry points for hacker-attacks to safety-critical features and thus new high-profile risks. AV with systems, operating precisely in a dynamic environment including unpredictable behavior of others, represent an unprecedented technological challenge and level of complexity. They need to be understood as entire system; test-driving in dedicated fields and real environments, in particular, needs to be complemented by alternative ways and validated methods. Regarding release to public roads, two paradigms are pursued – “self-certification” or “type permission”. Autonomous driving will involve a multiplicity of parties including new decision-making entities like learning algorithms and will enable innovative collective mobility concepts with vanishing traditional boundaries. Besides purely technological aspects, environmental, legal, socio-political and ethical issues need to be highlighted. Potential risks are double-sided and systemic by nature; they are coined by high complexity, uncertainty and ambiguity and multiple influencing factors. A stepwise governance approach with enlarged stakeholder participation seems to be advisable. As part of the process the public needs to be better informed and sensitized; expected benefits and concerns need to be balanced, safety targets to be fixed and verified. As commonly agreed transparency and communication should play a central role within the process.