Enterprise Liability and Risk Governance in the Regulation of Artificial Intelligence

Abstract

The growing autonomy of artificial intelligence systems is reshaping long-standing assumptions about legal responsibility. Traditional tort frameworks were built around human actors and clear causal links. These assumptions weaken when harm is produced by systems that learn, adapt and operate with limited human oversight. The paper examines how AI challenges established legal reasoning by analysing liability through two key dimensions. The first concerns the responsibility of human actors like developers, deployers and professional users, whose decisions shape AI behaviour. The second considers whether, and in what sense, liability can be attributed to AI systems themselves. The analysis also draws on concepts like enterprise liability, Ulrich Beck’s risk society theory and European Union’s risk based approach to AI regulation. The paper argues for a combined model that uses regulatory duties to prevent harm and tort law to address such harm. This framework aims to balance accountability, fairness and innovation.