China Unveils First Mandatory National Standard for L3/L4 Autonomous Driving, Effective July 2027

On June 16, 2026, the Ministry of Industry and Information Technology (MIIT) opened for public comment two draft mandatory national standards—“Safety Requirements for Automated Driving Systems of Intelligent Connected Vehicles” and another related standard—as well as the foreign-language version of the draft mandatory national standard “In-Vehicle Emergency Call System in Case of Accidents.” The public consultation period runs from June 17 to June 24, 2026. The standards are scheduled to take effect on July 1, 2027. As China’s first mandatory national standard specifically targeting Level 3 (conditional automation) and Level 4 (high automation) automated driving systems, its scope covers M-class (passenger-carrying) and N-class (goods-carrying) vehicles equipped with Level 3 and/or Level 4 automated driving systems, excluding automated parking systems. This standard represents a systematic upgrade from the voluntary national standard GB/T 44721–2024 issued in 2024, transitioning it from voluntary to mandatory. The standard stipulates that the safety performance of an Automated Driving System (ADS) must be no lower than that of a “qualified and attentive human driver” and must not pose unreasonable safety risks to users or other road users. When activated, the system must perform all dynamic driving tasks and continuously monitor whether it remains within its Operational Design Domain (ODD). Manufacturers must specify the conditions under which the system’s capabilities are achieved, the methods used for verification, and risk mitigation measures in case of system failure. The standard introduces a Safety Case mechanism, requiring manufacturers to systematically demonstrate the safety of their ADS using a structured “claim–argument–evidence” approach, covering capability boundaries, logical rationale, and test data. Simulation testing, closed-course testing, and on-road testing are integrated into a unified validation framework. Multi-source test results must jointly support safety claims in terms of scenario coverage and validation depth, while ensuring consistency between simulation toolchains and real-world vehicle test outcomes, thereby establishing a closed-loop, traceable safety assurance system. Recognizing the distinct operational characteristics of Level 3 and Level 4 systems, the standard sets differentiated requirements: For Level 3, emphasis is placed on human-machine handover processes, including monitoring the user’s readiness to take over, issuing timely alerts, and implementing risk-response mechanisms. If the user fails to assume control or if conditions triggering the Minimum Risk Maneuver (MRM) are met, the system must execute the MRM to bring the vehicle into a minimal-risk condition. For Level 4, the focus shifts to the system’s intrinsic ability to manage risks independently, without relying on remote assistance for dynamic driving tasks. Key elements—including ADS software, perception systems, automated driving functions, ODD definitions, and MRM strategies—are incorporated into the type approval assessment. New vehicle models applying for type approval must comply with the standard from its effective date. Existing approved models will be granted a 13-month transition period, with compliance required starting August 1, 2028. As of May 2026, over 57,000 kilometers of roads nationwide have been opened for intelligent connected vehicle testing, and 237 national and industry standards have been published.