Abstract
The automotive landscape is experiencing a paradigm shift, driven by the pervasive integration of artificial intelligence (AI) across all functional layers, the availability of previously unattainable data-processing capabilities, and an increasingly tight convergence of sensors, actuators, and advanced communication technologies. This perspective article explores the evolution from connected vehicles to Software (SW)-defined vehicles (SDVs), focusing on the emerging frontier of AI-defined vehicles (AIDVs) as a self-evolving architecture that integrates generative AI (GenAI) for scenario synthesis and agentic AI for autonomous decision-making. This article discusses the technology evolution and the role of AI and vehicle-to-everything (V2X) communication in enabling this transformation. The architecture of SDVs and the conceptual framework of AIDVs are examined, highlighting the transition from Hardware (HW)-centric to SW- and AI-centric designs. Recent advances and future directions are also discussed, presenting a multifaceted view of the opportunities and challenges by synthesising insights from research and industry trends and offering a forward-looking perspective on the technological, societal, and ethical factors shaping the evolution of automated intelligent transportation systems (ITSs).