Systemic explainability for autonomous driving: connecting ethics to law and society

Abstract

Abstract Self-driving vehicles challenge legal, socio-technical, and ethical frameworks due to their opacity, unpredictability, and autonomous decision-making, raising concerns for safety, trustworthiness, and responsibility attribution. In response, explainable artificial intelligence has been proposed to enhance transparency and trust in self-driving vehicles. This paper examines the role of explainability across the autonomous driving lifecycle, from development to deployment, through a thematic analysis of key dimensions, including safety, trust, transparency, trustworthiness, liability, and accountability. The analysis is carried out within the scope of European Union legal and ethical works, and shows how explainability supports functional and interactive safety, enables causal reconstruction of autonomous decisions, mitigates responsibility gaps, and fosters calibrated trust among users and stakeholders. Beyond these individual effects, the paper advances a systemic perspective on explainability, arguing that XAI operates as a structural enabler that reconnects technical choices to legal responsibility, ethical design, and social legitimacy. The principle influences autonomous driving from development throughout deployment, where explainability contributes to making autonomous driving systems intelligible, governable, responsible, and acceptable. The paper concludes that explainability should be understood not only as a technical tool, but also as a foundational and systemic design principle for autonomous vehicles.