As artificial intelligence (‘AI’) becomes pervasive, it is inevitable that there will be a corresponding increase in AI disputes. However, it is often no easy task to determine liability in AI-related disputes due to the unique features of many AI systems, including their complexity, opacity, and the ability of some AI systems to self-adapt. This paper considers the existing EU liability framework and concludes that, at present, it is not adequate to ensure that injured parties can be compensated for damage caused by AI. The argument presented is that Product Liability Directive 85/374/EEC should be amended to provide greater protection for claimants that suffer harm caused by AI, and proposals for revision are suggested. In addition, this paper argues that there are cogent reasons to design sector specific legislation and identify ex ante which party should be held strictly liable for certain classes of AI.