China’s Autonomous Vehicle Regulations

    By Andrew Zhang, Alice Zhu and Jason Xu, Fangda Partners
    0
    371
    Whatsapp
    Copy link

    MAIN

    TAIWAN

    The new energy vehicle (“NEV”) sector in mainland China has developed rapidly to become a large-scale and competitive industry. With recent advances in electrification, intelligence and connectivity in NEVs, rapid advances are now being made in autonomous driving technology, prompting ongoing legal debate and legislative action in response to this fast-evolving sector. Due to the inherently conservative nature of the law, current higher-level legislation (such as the Road Traffic Safety Law and the Product Quality Law) is still primarily premised on traditional, non-autonomous driving. However, there is considerable legislative and regulatory activity happening at the lower levels of the legal system – e.g., local regulations, technical standards, and normative documents – with numerous pilot projects and proactive exploration of ways of regulating autonomous driving.

    Andrew-Zhang
    Andrew Zhang
    Partner
    Fangda Partners
    Shanghai
    Tel: +86 21 2208 1152
    Email: andrew.zhang@fangdalaw.com

    Not surprisingly, technical requirements play a crucial role in the regulatory system in relation to autonomous driving. China’s regulatory framework for autonomous driving exhibits a clear three-tier structure:

      1. Higher-level laws and traditional legislation, which provide the fundamental legal principles;
      2. Normative documents, issued by authorities such as the National Development and Reform Commission (“NDRC”), the Ministry of Industry and Information Technology (“MIIT”) and the Ministry of Transport, which play a key guiding role in core areas such as market access, testing and data security; and
      3. Technical standards – including national, industry and local standards – which set out specific technical parameters and operational norms.

    The national technical standard governing autonomous driving (the “Classification of Driving Automation for Vehicles” (GB/T 40429-2021)) classifies autonomous driving technology into various levels. Levels L0 to L2 refer to “assisted driving” systems with more limited degrees of automation. For Levels L3 (conditional autonomous driving) and above, the system can perform the entire dynamic driving task in specific scenarios when activated, though human involvement is still required for L3, which is the current focus of regulatory activity and industry breakthroughs. This article provides a review and analysis of the legal and regulatory framework as it pertains to intelligent driving vehicles (i.e., autonomous vehicles) falling within levels L3 and above.

    Comprehensive Policies

    Alice Zhu
    Alice Zhu
    Partner
    Fangda Partners
    Shanghai
    Tel: +86 138 1779 2425; +852 6302 7363
    Email: alice.zhu@fangdalaw.com

    National level. In 2020, the NDRC and 10 other ministries issued the Strategy for the Innovative Development of Intelligent Vehicles, which clearly established the goal of achieving mass production of conditionally autonomous (L3) vehicles by 2025, whilst aiming for more use of highly autonomous (L4) vehicles in market oriented applications. In January 2024, the MIIT and four other ministries jointly issued the Notice on Launching Pilot Projects for “Vehicle-Road-Cloud Integration” Applications of Intelligent Connected Vehicles. This notice promotes unmanned and commercial applications of autonomous driving technology at levels L4 and below, and simultaneously launched pilot projects in 20 cities including Beijing and Shanghai. The continuously updated Guidelines for the Construction of the National Internet of Vehicles Industry Standard System establishes a comprehensive and systematic set of standards for intelligent connected vehicles, covering basic and key technology, product application, and testing and demonstration, providing a standardised basis for technological R&D and product access. In 2025, the MIIT released the Key Points of Automotive Standardisation Work, proposing the formulation of mandatory standards such as “safety requirements for autonomous driving systems” and the establishment of related standards.

    Local level. The Beijing Autonomous Vehicle Regulations issued in 2024 provide a clear legal basis and operational rules for high-level autonomous vehicle road testing, demonstration uses and commercial operations in Beijing. The Shenzhen Special Economic Zone Intelligent Connected Vehicle Administrative Regulations (“Shenzhen Regulations”) issued in 2022 comprehensively regulate registration for market access, on-road operation, vehicle-road collaboration, data management, traffic accident handling and legal liability for intelligent connected vehicles. The Regulations on Promoting the Innovative Application of Driverless Intelligent Connected Vehicles in Shanghai Pudong New Area and its implementation rules, issued in 2022, focus on the innovative application (testing, demonstration operation, commercial operation) of L4 driverless vehicles. In 2024, a limited number of “fully driverless” licences were issued and dedicated high-speed test roads were opened.

    Vehicle Access Qualifications

    Jason Xu
    Jason Xu
    Partner
    Fangda Partners
    Shanghai
    Tel: +86 21 2208 1157
    Email: jason.xu@fangdalaw.com

    China adopts a strict regulatory model for the use of autonomous vehicles on public roads. All mass-produced vehicles must obtain Road Motor Vehicle Manufacturer and Product Admission from the MIIT before being driven on public roads.

    Current status (L3 and above). At present, there are no unified, specific national access standards formally issued for mass-produced vehicles possessing autonomous driving functions at levels L3 and above. The core basis for access remains the Administrative Regulation on Road Motor Vehicle Manufacturer and Product Admission issued by the MIIT and related technical documents.

    Pilot mechanism. Currently, access by L3 or L4 vehicles to public roads is largely enabled through local pilot rules. For example, local regulations and implementing rules in Beijing, Shanghai and Shenzhen have established access conditions and approval processes for road testing, demonstration uses and even commercial operations within specific areas of their administrative regions. Vehicles entering these pilot areas must comply with the access standards set by the city.

    Based on the experience of local pilot projects, industry players expect that national access standards for L3 and above are currently in the pipeline. After being optimised through local pilot projects, the core requirements contained in existing pilot standards (especially those concerning function safety (fail-operational safety mechanisms), human-machine interaction, data recording, etc.), are likely to become important components of future standards.

    Accident and Product Liability

    The key regulatory focus with traffic accident liability is on clarifying the responsible party and the rules for determining liability. The Road Traffic Safety Law and its supporting regulations are mainly premised on the human-driven model, leaving significant gaps in the law when it comes to the allocation of liability in the context of self-driving or autonomous driving systems.

    Under the Shenzhen Regulations, when the L3 autonomous driving system is activated, the driver must respond immediately when the autonomous system requests human takeover. If a traffic accident occurs with a human-driven intelligent connected vehicle and the vehicle is at fault, the driver bears liability for compensation. For L4 and L5 driverless vehicles, the vehicle owner or operation manager (registered operation entity of the vehicle) bears primary liability.

    In addition, where defects in autonomous vehicles (which are considered products under PRC product liability laws), cause accidents, liability is determined based on the strict liability principle established in the Product Quality Law and the Civil Code. The producer is liable for damages caused by the defective products, and the consumer or victim only needs to prove the product defect, the damage and the causal relationship between the two. As mentioned above, local regulations on road traffic safety (such as article 54 of the Shenzhen Regulations) have clarified that if damage is caused by vehicle defects, the driver/owner/manager who has fulfilled its compensation obligations has the right to seek compensation from the producer, providing a practical path for applying the Product Quality Law. Local legislation mainly clarifies the triggering mechanism of the liability chain (initial liability allocation to frontline entities (driver/owner/operator), followed by defect-based recourse against manufacturers), without changing the basic attribution principle established by the Product Quality Law.

    Cybersecurity and Data Protection

    Autonomous vehicles rely heavily on network connectivity and data processing, making cybersecurity and data protection particularly critical. There are no dedicated regulations for data security protection in relation to autonomous vehicles at the local level, with these issues thus governed by rules and regulations at the national level. The Cybersecurity Law, Data Security Law and Personal Information Protection Law form the basic legal framework at the higher level. The Provisions on the Management of Automotive Data Security (Trial) issued in 2021 and the Notice on Strengthening the Management of Mapping and Geographic Information Security for Intelligent Connected Vehicles issued in July 2024 lay the regulatory foundation for data security of intelligent connected vehicles.

    Data collection and processing entities must comply with a series of requirements, such as storing key data (including geographic information) within China, and obtaining approval or passing security assessments for cross-border transfers of data. Enterprises must follow the “minimum necessary” principle when collecting data, and implement encryption and access controls for sensitive information. Processing important data requires risk assessment and submission of reports. In the event of a data security incident, emergency response plans must be activated and reported immediately.

    In addition, automobile companies must implement a cybersecurity graded protection system, build protection systems, monitor and block cyberattacks in real time, update software and patch vulnerabilities promptly, and strengthen in-vehicle software security testing and real-time monitoring of network status during operation.

    In conclusion, China’s legal and regulatory framework for autonomous vehicles is moving from fragmented pilot projects towards the establishment of a more systematic national regime. In the future, with unified standards, improved liability mechanisms and strengthened

    data regulation, autonomous vehicles at levels L3 and above are expected to enjoyclearer pathways for market access and a more comprehensive and predictable compliance regime at the national level. This will support the safe, healthy and sustainable development of the autonomous driving industry.

    Fangda Partners

    Fangda Partners
    Shanghai Office
    24/F, HKRI Centre Two, HKRI Taikoo Hui 288

    Shi Men Yi Road Shanghai 200041, China
    Tel: +86 21 2208 1166
    Fax: +86 21 5298 5599
    Email: email@fangdalaw.com

    Whatsapp
    Copy link