A review of technology evolution and risk of autonomous vehicles

The development of autonomous vehicles (AVs) have become a prominent research topic; however, many survey studies focus either on enabling technologies or on isolated risk issues. This approach therefore provides limited insight into how both dimensions of AV development evolve together. The present study employs scientometric analysis of Scopus-indexed journal articles to map the knowledge base of AV technology evolution and risk. The results of the study highlight influential documents and productive countries, identify major research clusters using the log-likelihood ratio (LLR), and reveal thematic shifts across three periods (early, middle, and late). To strengthen the analytical contribution, the revised manuscript synthesizes the interaction between technology phases, dominant methods, associated risks, and corresponding research responses. The findings indicate that early AV studies emphasized autonomy and dynamics. However, there was a subsequent shift towards systems, control, and learning. Moreover, there has been an increasing convergence with risk themes such as cybersecurity, safety assessment, and anomaly detection. These findings of this study offer a more integrated understanding of the co-evolution of AV research and indicate priority challenges for the safe deployment of learning-based methods. Ultimately, the insights provided in this review offer a valuable foundation for policymakers, automotive engineers, and researchers to develop holistic strategies that concurrently address technical innovations and their associated safety or regulatory risks.