[Chengze Observation • Platform Economy 40 Evaluation 26] Huang Yiping: How to share big data?

Zhongxin Jingwei August 2: How to share big data?

The author Huang Yiping Peking University National Development Research Institute, deputy dean, director of Peking University Digital Financial Research Center

Introduction: For data elements, we cannot simply apply the method of governance of traditional production factors to achieve big data sharing. It should also be mainly shared the results of big data analysis, rather than sharing the original data itself.

In April 2022, I presided over a discussion on the digital economy during the Boao Forum for Asia. At the meeting, the vice chairman of the Boao Forum for Asia and the chief representative of the Chinese side, Zhou Xiaochuan, president of the Chinese Society of Finance, proposed to overcome the international room with security algorithms. The contradiction of data is difficult to completely flow. If he borrows him, it can be understood as "the data does not exit, but the analysis results can be exited." This idea has great inspiration to me. At present, European and American countries are actively promoting the formulation of international digital trade rules, and put forward the claim of data cross -border free flow. However, this constitutes a challenge for other countries, including China. If you cannot flow freely based on national security and other considerations, you may restrict these countries to participate in international digital trade. Zhou Xiaochuan's thinking provides a solution for resolving this contradiction, that is, replacing the flow of original data with the flow of analysis results. This idea also opened my thinking space for domestic big data sharing.

"Data is the oil of the digital economy era." In reality, there are many successful cases of big data applications, including the precise marketing of products, personalized solutions for services, and effective evaluation of credit risk. Big data analysis can help improve economic efficiency, improve user experience, expand business scale, and promote economic and social innovation and development. Big data has become a very important production factors in the current economy. So far, most of the big data has been produced on the big technology platform, so large technology companies are the main practitioners of big data analysis. But if big data can be shared, the dividend of big data analysis will be greater. It may be for this consideration that the government has promoted data sharing in some fields in recent years. The purpose of setting up several big data credit reporting companies should be to allow more commercial banks to use big data credit risk assessment methods.

But as a special production factors, not all big data can be shared. How to achieve the effect of data sharing on this premise does need to be innovative. Discuss data sharing, first of all, you must face the issue of data ownership, who is the data of the data? Traditional production factors such as the effective use of capital and land are based on a prerequisite, which is to clearly define property rights. However, it is difficult to simply apply big data for this approach. The direct reason is that the ownership of some big data is difficult to be clearly defined. For example, the data of users on the platform and watching short videos on the platform not only contain some personal information, but also have online activities supported by the platform. Obviously, in this example, it is very difficult to draw the ownership line of data between users and platforms. A reasonable solution is that data involving personal characteristics is owned by individuals. The digital footprint left by users on the platform can be owned by individuals and platforms.

In real life, data confirmation also faces a challenge, that is, a balance between the protection of equity and the efficiency of use. The practice of China and the European Union provides two corresponding cases. In the past, China lacked effective data governance policies, neither confirmed, and insufficient protection. In this way, on the one hand, innovative activities using big data analysis are very active, but on the other hand, various violations and illegal behaviors are also very common. As a result, the government has taken many measures in recent years to increase data protection. The European Union has been doing earlier and better in terms of data protection. The shortcomings of the United States and China are due to more restrictions. Economic innovation based on big data analysis is relatively not active. Therefore, data governance, especially the confirmation, needs to adhere to one principle, is to protect both rights and innovation. Pay special attention to the contribution and rights of the platform for data confirmation.

Discussion data sharing also needs to face the problem of suitableness. What data can be shared and what data cannot be shared? In principle, at least three types of data cannot be shared: one is personal privacy, the other is commercial secrets, and the third is national security. Individual names, age, gender, education level, family address, contact information, etc. are all personal privacy. They should be clearly belonging to individuals, and of course they cannot be easily shared. The common practice in the industry is to desensitize behavior and transaction data, and analysts cannot go back to specific individuals or people. Similarly, if it is about commercial secrets and national security, data cannot be disclosed casually. It is not difficult to reach consensus on these large principles. It is difficult to grasp how to grasp during the execution process. Taking national security as an example, if the scope is widened, it may be possible to determine that most of the data has a certain sensitivity. Changes in the way of crowd. Therefore, how to properly grasp this degree is also a great test. If the standard is not strict enough, many hidden dangers may be caused. But if the standard is too strict, big data analysis will not start.

Discuss data sharing, and the question to be answered in the end is how to share it? Big data is actually very suitable for sharing, that is, it can be copied unlimited. This is a fundamental difference between data factor and traditional production factors. Whether it is capital or land, if there is already a company in use, other companies cannot use them at the same time. However, big data is different. If it is needed, a set of data can be used by countless companies at the same time. Perhaps this is why big data analysis can produce extraordinary returns. But this advantage will also cause a disadvantage, that is, how the data owner guarantees that the buyer will not copy the data for others. If it cannot be guaranteed, then the owner can only sell it once. Complete internalization. In this way, it is difficult for an owner to have sufficient incentives and resources to collect, clean up, and analyze the time -consuming and labor -intensive collection. If the big data is not produced, the sharing will not be able to talk about it. But there are already examples of public data sharing in real life. The so -called public data mainly refers to a large amount of static data accumulated by the public sector. Typical examples are taxation, social security, judicial, and even hydropower. These data have been formed, and no additional investment is required. At present, some areas have established local data platforms, integrated existing public data, supported commercial banks' credit risk assessment, and provided loans for SMEs, and achieved good results. Guangdong Province, Zhejiang Province, Suzhou, and Zibo City have different platform forms that have different attempts, the leading agencies are different, and even the data covering the data is very different. Data, support inclusive financial business.

In particular, it is necessary to point out that even such a platform for shared public data, there is no raw data. The function of the platform only provides an interface, allowing the authorized bank algorithm tools to enter different public database for operations and get results. In other words, these financial information service platforms are essentially the idea proposed by Zhou Xiaochuan. "The original data is not the system, but the analysis results can be released." "Public data" cannot go out of the system, considers the rights and interests, and also has safe consideration. But this reason is even more prominent on the big data on the large technology platform. Because most public data are static, when do you pay and how much water and electricity costs are paid? It is used in the system or outside the system, which is not much different. But big data is dynamic. If you leave the system after division, it may be difficult to produce the same big data analysis effect.

Recently, there is a view that since big data is a new production factors, it is important and sensitive, so it is best to be mastered by government departments or state -owned enterprises. This idea has a certain reason, because in reality, there are indeed many non -compliance and illegal behaviors in data processing. Of course, almost all large technology companies are private enterprises, which is also a fact. However, the effective way to solve the legal problem of compliance is to strengthen supervision, rather than set the data to the state -owned sector. How can data be concentrated in the state -owned department, this technical problem is not easy to solve. Even if it can, the state -owned sector's behavior is relatively high, but the innovation is generally relatively low, so the vitality of big data analysis is lost. The best and even the only solution is to strengthen the supervision of private technology companies. In fact, in the past two years, the laws and systems of data protection have been introduced frequently, and the data protection awareness of large technology companies has been greatly strengthened. Even if scholars engage in academic research, the difficulty of obtaining data has increased significantly.

In short, do not apply the management methods of traditional production factors to manage data production factors. Whether it is confirmation or sharing, it is necessary to adapt to the data characteristics. The fundamental purpose is to achieve a balance between protecting the rights and interests of relevant parties and exerting big data effectiveness. Big data sharing is a direction worthy of longing and hard work, but in the implementation, we must avoid simple and rude approach, especially for sharing data with everyone. In fact, many original data cannot be taken out. Among them, there are considerations of personal privacy, commercial secrets, and national security. It is also because the cost of data replication is very low. More importantly, many data Once a lot of data leaves its native platform system. Its value will be greatly discounted or even worthless. Therefore, the reasonable approach to promote sharing is to establish a data sharing platform. The organization that needs to use big data can use the access port provided by the platform to use the original data of the large technology platform for operation, and then output the result, that is, "the original data does not out of the system, the operation is the operation, the operation is the operation, the operation is the operation. Results out of the system. " This is the inspiration I have obtained from Zhou Xiaochuan's thinking on the Boao Forum for Asia. (Zhongxin Jingwei APP)

This article was selected by the Sino -Singapore Jingwei Research Institute. The copy of the work produced by the selected work, the copyright of the work, is not authorized by any unit or individual. The views involved in the selected content only represent the original author and do not represent the view of the Sino -Singapore Jingwei.

Editor in charge: Song Yafen

[Chengze Observation · Platform Economy 40 Evaluation Twenty -five] Shen Yan: There is a lot of promotion of algorithm audits

[Chengze Observation · Platform Economy 40 Twenty -four Evaluation] Zhang Junni: Active improvement of algorithm transparency is conducive to the sustainable development of platform enterprises

[Chengze Observation · Platform Economy 40 Evaluation Twenty -three] Yu Changhua: Does the digital economy improve the investment efficiency?

Pay attention to the official WeChat public account of JWVIEW (JWVIEW) to get more elite financial information.

- END -