Summary of Key Points
With the widespread adoption of new energy vehicles, data is scattered among various entities such as car manufacturers, power grids, and testing institutions, creating what can be referred to as "information islands." This leads to difficulties in determining responsibility for autonomous driving accidents, a lack of battery safety warnings, and inaccurate insurance pricing. The State-owned Assets Supervision and Administration Commission (SASAC) has led four central state-owned enterprises—China National Automotive Testing & Research Center (CNATR), Changan Automobile Group, State Grid Corporation of China, and China Telecom—to pilot the establishment of a "credible data space." This initiative uses technologies like privacy computing to ensure that data remains within its respective entities while still being usable. The goal is to address these issues. However, the pilot faces practical challenges such as data ownership clarification, lack of unified standards, and long-term operational sustainability. If successful, it could serve as a model for digital infrastructure development and be replicated in other industries.
Data "Islands": Three Major Challenges with New Energy Vehicles
The data from new energy vehicles is currently like documents scattered in different drawers, with no entity possessing the complete picture:
1. Difficulty in Determining Responsibility for Autonomous Driving Accidents: After a collision, car manufacturers claim the owner did not take control of the vehicle, while owners argue that the system failed to issue a warning. Insurance companies often deny claims or offer lower payouts due to unclear responsibility, as the autonomous driving logs are held by the manufacturers and there is no neutral data to serve as evidence.
2. Significant Battery Safety Risks: Before a battery overheats and catches fire, there are signs such as abnormal voltage/temperature readings and changes in charging patterns. However, more than half of this data is held by the power grid (State Grid Corporation of China), and the remaining half by the car manufacturers. Without access to all the data, potential problem vehicles can continue to be on the road.
3. Imprecise After-sales Services: The valuation of used cars relies on experience (e.g., appearance), and insurance pricing is based on a one-size-fits-all approach for vehicle models. Without comprehensive data on driving behavior, charging history, and maintenance records, it is impossible to accurately assess risks.
Credible Data Space: How to Achieve Shared Data Without Breaches of Privacy?
This system functions like an "encrypted shared office," with each of the four central state-owned enterprises managing a portion of the data:
- China Telecom: Provides the technical infrastructure, including network connectivity, computing power, privacy computing tools, and encrypted channels to ensure data security.
- Changan Automobile Group: Supplies core vehicle data, such as autonomous driving logs and driver operation records, which are crucial for accident liability determination.
- State Grid Corporation of China: Shares charging data; over 60% of battery health information comes from this source, which can help predict the risk of spontaneous combustion.
- China National Automotive Testing & Research Center (CNATR): Acts as a neutral party to validate data and calculation models, ensuring the accuracy and credibility of the results.
Operation Process: For example, if an insurance company wants to assess the risk of a particular vehicle, it submits a request to the system. China Telecom then sends instructions to the other three companies, which process the data on their servers (cleaning and extracting relevant features) before transmitting only encrypted intermediate results. The final report (e.g., "The vehicle's autonomous driving system was functioning normally, and the owner did not take control in time") is generated without exposing any original data.
Practical Issues That Can Be Solved
This approach can address several key problems:
1. Consistent Accountability in Autonomous Driving Accidents: Multiple dimensions of data (autonomous driving status, driver behavior, charging records, maintenance history) can be reviewed after an accident, generating a report recognized by the judiciary. This eliminates the need for lengthy evidence collection and reduces unfair blame分配.
2. Early Battery Safety Warnings: By combining data from all four entities, potential battery safety issues can be identified, preventing fires before they occur.
3. Improved After-sales Services:
- Insurance: Premiums can be tailored based on driving behavior and charging habits (e.g., lower for vehicles with stable performance and proper charging).
- Used Cars: More accurate valuations are possible due to comprehensive data, including battery health, which directly affects the vehicle's value.
- Regulation: Real-time monitoring of battery safety can help prevent accidents.
Challenges Faced by the Pilot
Although the concept is promising, there are several practical obstacles:
1. Data Ownership and Profit Distribution: Data is a valuable asset for these central state-owned enterprises. There are no clear rules regarding how profits generated from using this data should be shared among them, nor are the terms of privacy authorization and liability in case of breaches defined.
2. Lack of Unified Standards: Different car manufacturers and power grids use different data formats, requiring significant investment to adapt their systems. Additionally, unified privacy computing models are needed, which poses a technical challenge.
3. Sustainability of the Pilot: The pilot program is only scheduled for one year. After its completion, there are uncertainties regarding who will be responsible for maintaining the system and where the funding will come from. Without resolving these issues, it will be difficult to scale the initiative across the entire industry.
Significance of the Pilot
This project marks the first implementation of the national "Credible Data Space Development Action Plan" and has far-reaching implications beyond the automotive sector:
- Industry Model: If successful in the automotive industry, this model can be replicated in other sectors such as energy, high-end equipment, and healthcare (e.g., sharing clinical trial data).
- Transformation of the Automotive After-sales Market: More fair accident liability determination, more reasonable insurance pricing, and more transparent used car valuations will enhance industry efficiency.
- Pioneering a New Model: It explores whether a collaborative and profit-sharing approach can break down data islands and establish a new digital order.
In summary, the "credible data space" represents a potential solution to the challenges associated with new energy vehicle data. However, success depends on resolving issues related to data ownership, standards, and operational sustainability. After the one-year pilot, we will have a better understanding of whether this approach can truly be adopted on a wider scale, revealing the "truth" hidden within the data.