China Unveils Mandatory National Standard for Combined Driving Assistance Systems, Effective January 2027
New national compulsory standard GB47955—2026 sets unified safety benchmarks for L2 intelligent driving functions amid soaring market penetration, addressing widespread industry operational and consumer protection challenges
New national mandatory standard GB47955—2026, formally titled Intelligent Connected Vehicles – Safety Requirements for Combined Driving Assistance Systems, has secured official approval from the State Administration for Market Regulation and the Standardisation Administration of China. Drafted under the administration of the Ministry of Industry and Information Technology, the full regulatory text will enter into force on 1 January 2027.
Market uptake of passenger cars fitted with combined driving assistance has climbed rapidly across China, with intelligent driving hardware moving from premium optional extras to standard fitments on affordable family vehicles. The newly issued binding standard establishes clear safety thresholds for all L2-class vehicles, forming robust safeguards for road mobility powered by smart automotive technology.
Industry-wide adoption of assisted driving has reshaped consumer vehicle markets as technical upgrades broaden access to automated driving functions. Industry data collated this year shows seven out of every ten new passenger cars sold in the domestic market carry combined driving assistance hardware, while high-level navigation on autopilot (NOA) functions for highways and urban roads have seen their installation rate surpass 30 per cent.
Vehicles priced from 100,000 yuan entry-level models through to mainstream new energy vehicles priced between 200,000 and 300,000 yuan now almost universally come equipped with L2 combined driving assistance as standard. Lane centring, adaptive cruise control, automatic lane change and navigation-assisted driving have become daily-use features for motorists, stripping intelligent driving of its former status as a luxury-only feature.
Widespread rollout of these systems has coincided with mounting structural obstacles that hinder sustainable industry growth. Without enforceable unified rules, inconsistent safety performance, rising road incidents and mounting consumer disputes have become persistent sector-wide hurdles. Misleading product positioning and overblown marketing claims remain prevalent across the supply chain. Many vehicle manufacturers and retail outlets deliberately blur the dividing line between driver-assist technology and fully autonomous driving, labelling L2 combined assistance as full self-driving capability. Such promotional language downplays the statutory rule that drivers must maintain constant observation and vehicle control at all times, misleading vehicle owners into lowering their vigilance behind the wheel.
Misguided consumer perceptions fostered by exaggerated advertising have triggered risky road behaviour. Many motorists remove both hands from steering wheels, scroll through mobile devices or take their eyes off carriageways for extended periods while NOA functions are active, creating substantial hazards on motorways and arterial routes. The Intelligent Driving Safety White Paper, published by China Automotive Technology and Research Center, records a more than 200 per cent year-on-year rise in traffic accidents between 2022 and 2025 linked to overreliance on driving assistance hardware and neglect of personal driving responsibilities. Misunderstanding of system limits ranks as the primary trigger for smart driving-related collisions.
Disparate safety standards and uneven product quality represent a further prominent industry issue. When guidance was limited to voluntary recommended specifications, the sector lacked uniform testing protocols and baseline market access criteria. Some manufacturers cut production costs by reducing radar and camera sensor hardware, alongside simplified calibration algorithms for obstacle detection and automatic emergency braking (AEB). Such cost-cutting measures create tangible safety risks when vehicles operate under adverse conditions, including heavy rainfall, backlit sunlight, night-time travel, roadwork zones and mixed traffic featuring non-motorised vehicles. Detection failures, delayed braking responses and inactive lane departure warnings occur frequently in these complex scenarios. Vehicles within identical price brackets deliver vastly inconsistent intelligent driving safety performance, leaving buyers unable to accurately assess real-world functionality and exposing consumers to underperforming products with impressive surface technical parameters.
Long-standing structural issues also persist around human-machine interface design and evidence collection following road collisions. Most vehicle manufacturers only print small-font risk disclaimers within owner handbooks, while in-vehicle alerts for distracted or hands-off driving remain infrequent and weak, failing to deliver meaningful driver intervention. When traffic collisions occur, vehicle manufacturers retain full access to real-time operational logs, system decision records and warning trigger data, yet private vehicle owners hold no formal channels to retrieve such evidence. Ambiguous liability frameworks stemming from this information imbalance create a backlog of unresolved consumer disputes and raise barriers for claimants seeking redress.
A specialist from the brand think tank affiliated with Yunnan Administration for Market Regulation shared insights with China Consumer News on the unregulated state of the domestic combined driving assistance market prior to the new standard’s launch. Manufacturers operated under self-devised internal rules, with no consistent naming conventions for system functions, clear capability boundaries or standardised safety redundancy design frameworks. Cumulative systemic safety risks have built up as penetration of intelligent driving vehicles continues to expand. The new compulsory national standard does not restrict technological innovation within the automotive sector; instead, it delivers universal baseline regulations to eliminate the widespread misconception that drivers may disengage supervision while driving assistance functions operate, and restructures orderly market conduct across the industry.
The new standard draws on the full spectrum of China’s complex, diverse road traffic conditions, accommodates divergent technical pathways adopted by vehicle makers, and integrates optimisations tailored to local transport environments after benchmarking leading international vehicle regulations. It introduces dedicated testing scenarios unique to domestic road networks, including hazard avoidance through construction zones and compatibility with mixed traffic containing bicycles and electric two-wheelers. A complete set of quantifiable safety metrics covers product classification, functional safety and human-vehicle coordination. All new vehicle models submitted for type approval from January 2027 must satisfy every clause of the standard; any vehicle failing compliance testing faces a ban on production and retail distribution.
Formal unified terminology for combined driving assistance systems is established within the document, which codifies the core operational principle: such technology only delivers lateral and longitudinal vehicle control support under the continuous supervision and physical control of a human driver, drawing an unambiguous distinction between assistance tools and fully driverless autonomous vehicles. All available combined driving assistance products are split into three defined categories: basic single-lane cruise control systems, basic multi-lane cruise control systems, and navigation-assisted driving systems. Differentiated, targeted safety assessment criteria apply to each tier, enabling consumers to clearly identify functional limitations and eliminate ambiguous, overstated marketing messaging.
To curb misleading promotional practices, the standard mandates full disclosure of each intelligent driving system’s permitted road types, operational speed thresholds, environmental constraints, functional limitations and prohibited usage scenarios within official product brochures and new vehicle handover training sessions. Marketing materials cannot claim operational capability beyond defined system boundaries. Models engineered solely for highway NOA operation are prohibited from advertising compatibility with all urban road networks, while basic single-lane assistance vehicles cannot market automatic lane change or intelligent overtaking navigation functions.
Regulations governing driver monitoring form a core pillar of the new standard’s safety architecture, introduced to mitigate collisions stemming from distracted, hands-free driving. Every combined driving assistance system must integrate dual monitoring modules to track hand contact with the steering wheel and continuous forward eye-line observation. Once the hardware detects sustained hands-off steering or diverted driver sightlines without timely human intervention, vehicles must activate tiered risk mitigation protocols. These include audio-visual warnings, active speed reduction and emergency stopping where necessary. The intelligent driving function will lock and remain unavailable for reactivation throughout the current driving cycle, closing technical loopholes that enable unsafe misuse of assistance technology.
Detailed specifications also formalise clear Operational Design Domain (ODD) boundaries, outlining permitted road classifications, light intensity thresholds, rain and fog density limits and authorised traffic environments. Manufacturers are required to list all complex road conditions where systems cannot operate reliably within vehicle handbooks, ensuring full transparency around functional limitations and eliminating vague, undisclosed operating constraints.
The specialist from Yunnan market regulation authorities noted that the standard resolves long-running ambiguity around functional boundaries through quantified rules governing operational environments, human-machine interaction and failure response mechanisms. Mandatory comprehensive testing regimes eliminate the practice of marketing hypothetical laboratory performance as universal real-world capability, while screening out low-specification, marketing-heavy products with inadequate safety design. All vehicle manufacturers will face pressure to upgrade end-to-end safety redundancy architecture. The standard establishes minimum industry safety benchmarks, and consistent certification enforcement, routine market supervision and ongoing consumer safety education will sustain long-term regulatory outcomes.
For private motorists, the enforced standard delivers two immediate tangible benefits: improved baseline safety for mass-market intelligent vehicles and streamlined procedures for consumer dispute resolution. Multi-layered track testing and real-world road verification frameworks cover frequent faults including AEB failure and uncommanded lane departure, requiring tens of thousands of simulation trials and full-condition on-road validation across urban carriageways, national highways, construction zones and mixed rural traffic corridors.
A special researcher at Su Commercial Bank outlined how these rigid testing obligations push automotive manufacturers to prioritise genuine technical performance over superficial marketing, reduce consumer misunderstanding of system functionality from the point of vehicle design, and cultivate a more secure, sustainable operating landscape for the whole automotive sector.
Long-standing obstacles around evidence collection and liability definition following intelligent driving collisions receive targeted regulatory fixes. Every new vehicle fitted with combined driving assistance hardware must carry dedicated intelligent driving data recording equipment, which encrypts and permanently stores full operational data generated while assistance functions are active. Recorded datasets cover vehicle speed, lane positioning and obstacle detection readings, with built-in safeguards preventing unilateral tampering or deletion of stored information. Regulators, vehicle manufacturers and individual vehicle owners gain lawful access to complete records post-collision, enabling objective reconstruction of incident sequences, accurate identification of root causes and balancing of the information asymmetry that previously disadvantaged claimants. The overall financial and procedural burden of consumer legal redress sees meaningful reduction.
Formalised protocols govern over-the-air (OTA) vehicle software updates under the new standard. Any major firmware revisions altering intelligent driving functionality, ODD parameters or safety logic must complete full safety validation and submit relevant documentation to regulatory bodies before public rollout. Manufacturers are forbidden to distribute untested updates carrying latent safety risks, delivering comprehensive protection for smart vehicle safety across full product lifecycles.
From a long-term industrial perspective, the standard fills the regulatory vacuum surrounding L2-class combined driving assistance in China, forming a complete tiered governance framework alongside upcoming standards for Level 3 and higher autonomous driving systems. Together, these documents create a holistic matrix of safety specifications covering all tiers of intelligent vehicle technology. Balanced provisions accommodate continuous technical iteration for automotive manufacturers while maintaining non-negotiable minimum standards for public road safety.
