The Notion of Defectiveness Applied to Autonomous Vehicles: The Need for New Liability Bases for Artificial Intelligence

Abstract

Both the US and the EU product liability regimes are based on the notion of defectiveness of the product. However, in the case of damages caused by autonomous vehicles, such notion proves to be profoundly inadequate for consumer protection. In fact, from the European perspective, the defectiveness of a product is assessed through the so-called consumer expectation test, according to which a product is defective when it does not provide the safety a person is entitled to expect. However, such approach is inadequate in the context of autonomous vehicles as it leads to unreasonably high safety expectations. By contrast, the US product liability doctrine adopts the so-called risk-utility test, according to which a product is defective if the foreseeable risks of harm could have been reduced or avoided by the adoption of a reasonable alternative design. Such approach is nonetheless undesirable as it links safety to market forces. This article aims at analyzing in comparative perspective the current legislation concerning damages caused by autonomous systems, with a view to devising new possible solutions and alternative approaches to product liability for Artificial Intelligence.