Liability for the Fact of Autonomous Artificial Intelligence Agents. Things, Agencies and Legal Actors
The possibility of widening producers’ or users’ strict liability for actions by artificial agents even when they are not fully predictable and that of extending the status of legal actor to some of these artificial agents so as to attribute the damaging fact directly to them are put in relation.
Abstract
Abstract European Union institutions are currently studying the regulatory intervention needed to face liability problems solicited by the use of autonomous artificial agents, whose unforeseeable actions may result in damage for their users or third parties. This paper intends to analyze some of the solutions proposed, by putting in relation – also from a historical perspective – the possibility of widening producers’ or users’ strict liability for actions by artificial agents even when they are not fully predictable and that of extending the status of legal actor to some of these artificial agents, so as to attribute the damaging fact directly to them.