Rethinking compensation in light of the development of AI

The opacity, autonomy and complexity of AI systems can stand in the way of a fair and efficient allocation of risk and loss. The European Commission (EC) has recognized this and has addressed these matters in two proposals for two directives: the AI Liability Directive and a new Product Liability Directive. Both Directives address information asymmetries between parties in a liability claim by providing new rules on the burden of proof. In addition, the proposed Product Liability Directive has been brought 'up-to-date' by explicitly incorporating new technical developments and by ending the debate on software as a product. Nevertheless, the injured party might still not receive compensation for the damage caused by a defective AI-system. This gives rise to the question of whether the legal framework for compensation should be revised more boldly to ensure compensation for the damage suffered by the injured party. This contribution will explore such a bold approach by delving into compensation funds. More specifically, this contribution will examine how a compensation fund for damage caused by AI-systems can be designed as well as what its boundaries could or should be and what its benefits could be.