Vicarious Liability: A Solution to a Problem of AI Responsibility

Abstract

Who is capable when a man-made intelligence machine makes something turn out badly? Or is there a void in the assignation of blame? The responses can either state that there is a single responsibility gap, multiple responsibility gaps, or none at all. The issue can be summarized as follows: On the one hand, it seems right to hold someone accountable for something that an AI machine did wrong; On the other hand, it doesn't appear that anyone deserves to be held accountable for this error. The study concentrates on a specific aspect of the AI responsibility gap in this article: In cases where AI machines have design flaws, it makes sense that someone should bear the legal costs; However, there does not appear to be such a suitable bearer. The study approaches this issue according to the legitimate point of view and propose vicarious responsibility of computer-based intelligence makers as an answer for this issue. Our proposition comes in two variations: The first one is more limited in scope, but it is simple to incorporate into existing legal frameworks; The second one can be used in a wider range of situations, but it requires legal frameworks to be updated. A broader definition of vicarious liability is used in the latter variant. Finally, study draw attention to the important insights that vicarious liability provides for closing the moral AI responsibility gap.