The Insurance Perspective on Prevention and Compensation Issues Relating to Damage Caused by Machines

This paper addresses the issue of automation coverage for costs in the event of damage caused by an automated decision-making process. It will consider civil liability and insurance from the point of view of problems related to the proof of a causal nexus between wrongdoing and losses. Starting from a study on causation, this paper focuses on liability and insurance in case of automation: their changing role in an automated world and various perspectives taking also into account recent European perspectives and developments. The thesis that the paper proposes is that legal liability is not a sufficient instrument to permit effective prevention and compensation in the case of damage caused by full algorithmic automation. This is particularly so because it could be not always possible to trace back to a specific human actor, as the European Commission underscored in its recommendation on civil law rules on robotics (2015/2103(INL)). Of course, legislators can intervene by reshaping the civil liability, for instance, by eliminating the proof of causal link or introducing new forms of strict liability. We intend to propose an alternative/ complementary way considering the role of insurance system, particularly liability insurance, which is generally intended as instrument to manage and transfer risks (both private companies and public funds) in compensating victims but also in preventing losses by educating the insured machines thanks to the data acquired. Il saggio è stato presentato in occasione del convegno Madrid 7 ottobre 2020 II Congreso Internacional de Derecho de Seguros/II International Congress on Insurance Law e pubblicato nel volume Dimensiones y desafìos del seguro de responsabilidad civil, Martinez Munos (ed), Madrid, 2021, isbn 9788413461526