Artificial Intelligence Systems and Liability – An Inquiry into the Tort of Negligence Paradigm

Abstract

With the rapid development of artificial intelligence systems in the last decade, legal liability arising out of the interaction between humans and artificial intelligence systems may become commonplace in the near future. While there are many academic opinions demanding for a new legal framework to address the new technology of artificial intelligence systems, this paper endeavours to draw from traditional tort law and build a realistic and workable structure of ascribing liability arising from interactions between humans and artificial intelligence systems. However, it is limited to the liability arising from negligence and hence primarily focuses on the tort of negligence. To this end, first, it analyses the application of the reasonable person test on matters involving interactions with artificial intelligence systems. Second, it deals with the application of previously known methods of calculating risk in human interactions with artificial intelligence systems. Third, it deals with defences against the tort of negligence claims and their applications on matters involving interactions with AI systems. And finally, it briefly looks into causation and how to determine if a negligence claim has merit or is trivial. The paper concludes that the liability in a negligence claim for an injury occurring during a human interaction with an artificial intelligence system will be attached to one or more of three parties: the developer, the manufacturer, and/or the user. Any other parties simply would not have enough causal link with the injury for it to be a legitimate claim.