Global Autonomous Driving Regulation to Shift to "Safety Case" Model by 2026

Recently, Li Jun, an academician of the Chinese Academy of Engineering, stated that 2026 will mark a structural turning point in global regulatory oversight of autonomous driving safety, with the focus shifting from compliance with predefined rules to "safety validation." Current regulations rely on preset functional checklists—automakers can launch vehicles as long as they meet these stipulations. However, this approach struggles to address the long-tail risks inherent in AI-driven systems and fails to adequately assess decision-making logic and risk-prediction capabilities in complex scenarios. Consequently, liability often remains ambiguous after accidents, leaving drivers as the primary victims. To address this issue, global regulatory trends are shifting from "post-incident accountability" to "pre-deployment validation," requiring companies to systematically demonstrate the safety of their autonomous driving systems through robust evidence. China plans to implement new standards on July 1, 2027, which will introduce an obligation for "in-service monitoring and reporting," clearly holding automakers accountable for vehicle safety throughout the entire product lifecycle. Under this new model, automakers must establish a comprehensive evidence chain covering design philosophy, testing and validation, and real-world operational data, thereby enhancing transparency in AI decision-making processes and improving regulatory traceability and user trust. This framework also compels companies to embed safety standards into every stage of product development—not merely relying on hardware specifications—to strengthen system performance in complex driving scenarios. Going forward, the ability to establish a credible "self-validation" system will become a critical competitive advantage for automakers in the commercialization race for autonomous driving.