Interpretation of the "Data Clause" of the Anti-Unfair Competition Act (revised in 2025)

Written by Mr Charles Feng, IP Expert and collaborator of the China IP SME Helpdesk, and Deng Sihan and Xue Lian.

In the era of digital economy, data has increasingly become an important factor of production for operators, and the acquisition and utilization of data resources is an extremely important resource allocation. Article 13, Paragraph 3 of the Anti-Unfair Competition Law (revised in 2025) adds a "data clause", which makes special provisions on unfair competition that infringes on data rights and interests, that is, operators shall not obtain or use data legally held by other operators by fraud, coercion, avoidance or destruction of technical management measures, etc., to harm the legitimate rights and interests of other operators and disrupt the order of market competition. The subject matter protected by this clause is "data lawfully held by other business operators". In order to avoid data ownership disputes arising from the diversification of data generation and user entities, the interest protected by this clause is the control and utilization interest of the operator over the commercially valuable data formed based on substantive investment, rather than the ownership of the data itself. This requires that "legally held data" has two elements: "lawful holding" and "competitive interests".

1. "Legal holding" and "competitive interests"

In judicial practice, the plaintiff must first prove that it enjoys stable and legal control over the data involved in the case, and that its claimed data rights and interests have a legitimacy basis, before the court can further consider whether the defendant's behavior constitutes unfair competition.1 Specifically, legal possession of data requires actual control over the data and legitimate data sources.

(1) Have the ability to control the data

Data possession requires both a de facto state of control over data and a willingness to dominate data. For example, it has the conditions and capabilities to collect, store, transmit, and arbitrarily retrieve user data through technical means, and sets certain protection measures for user data to exclude illegal interference from others.

(2) The data source is legal

The legal sources of data include legal provisions, contractual agreements and factual basis. "Legal holding" based on legal provisions refers to the holding and control of data by operators in accordance with the direct provisions or authorization of laws and administrative regulations. For example, the meteorological department collects, processes and publishes meteorological data in accordance with the Meteorological Law, and its possession of such data has an indisputable legal basis.2 The factual basis is mainly derived from the holding of input and control. This is often based on data collection through public channels or data held by operators.

Based on contractual agreements, operators acquire, use and control data within the agreed scope by signing an agreement with the data source party. Operators obtain data after being authorized by the entity that generates the data, and the typical situation is the user agreement or privacy policy signed between the network platform service provider and the network user. Such agreements should clearly inform users of the purpose, method and scope of collecting personal information, and follow the principles of "informed and consent" and "reasonable and necessary".

i. The principle of "informed and consent"

Article 1035 of the Civil Code stipulates that the consent of the natural person shall be obtained when processing personal information; Disclosure of processing rules; Clearly state the purpose, method and scope of processing information, and do not violate the provisions of laws and administrative regulations and the agreement between the two parties. Most of the information of online users is personal information, so online platform service providers should follow the principle of "informed and consent" to collect user information.

In the case of an unfair competition dispute between a technology company and a cultural media company,3 A technology company is the operator of APP A, and a cultural media company is the operator of APP B. There are 50,392 videos involved in the case on the B APP that are consistent with the short videos of A APP, and they contain the proprietary code of A APP. The video involved in the case contained 19,079 APP user nicknames, avatars and other registration information, of which 15,924 were the same as APP A; The content, order, and punctuation of the 127 comments are also the same as those of the A APP. The Haidian District Court of Beijing held that a technology company signed a user agreement with network users to collect and use users' personal information and text content of the video comments involved in the case with the consent of users, and the registration information such as nicknames and avatars of 19,079 APP users and the comment content of the videos involved constituted the data of Platform A.

ii. The principle of "reasonable and necessary"

Article 1035 of the Civil Code stipulates that the processing of personal information should follow the principles of legality, legitimacy, and necessity, so network service providers should follow the principle of necessity when collecting users' personal information. Article 13, Paragraph 1, Paragraph 2 of the Personal Information Protection Act provides an exception for processing personal information that is necessary for the conclusion and performance of a contract to which an individual is a party, and does not require the consent of the individual.

In the case of Luo and a software company over infringement of personal information rights and interests,4 When Luo registered on the website, the website required him to fill in user portrait information such as "occupation", "learning purpose", "school age" and "English proficiency". There are no options such as "skip" or "refuse" in the above process, nor are there any prompts for authorization and consent to the collection of personal information. A software company argued that collecting user profile information for automated decision-making is necessary to provide services and therefore does not require individual user consent.

The Beijing Internet Court held that if the absence of information processing would make it impossible to realize the basic functional services agreed in the contract or the additional functional services independently selected by the user, it could be determined that it was necessary for the conclusion and performance of the contract, otherwise it would not be recognized. Referring to the provisions of the "Provisions on the Scope of Necessary Personal Information for Common Types of Mobile Internet Applications" (Guoxinban Mi Zi [2021] No. 14) issued by the Cyberspace Administration of China, the Ministry of Industry and Information Technology, the Ministry of Public Security, and the State Administration for Market Regulation, the basic functional services of learning and education apps are "online tutoring, online classrooms, etc.", and the necessary personal information is the mobile phone number of registered users. The website involved in the case is an educational website, and its basic functions and services do not include information push to users through automated decision-making. Accordingly, the court found that a software company's collection of Luo's user portrait information without consent infringed on Luo's personal information rights and interests.

(3) Competitive interests

Data has competitive interests means that operators collect, sort, process and maintain data in accordance with the law through their own labor, capital, technology and other substantive investments, so as to form actual control of data and stable interest expectations. Protected data is usually a collection of data formed by operators investing a lot of manpower, material resources, financial resources, time and other costs.5 If the operator provides evidence to prove that it actually invests in the labor, technology or capital cost of the collection, storage and processing of relevant data, and the relevant data can bring actual or potential economic benefits to the operator in specific usage scenarios, it should generally be determined that the data can bring a competitive advantage to the platform.6

In the above-mentioned unfair competition dispute case between a technology company and a cultural media company,7 The Haidian District Court of Beijing held that a technology company, as the collector and controller of the overall short video of Platform A, paid huge costs in the process of collecting, storing, processing, transmitting, and presenting the overall short video, and enjoyed important business interests for the short video as a whole.

However, for the original data left by users on the online platform due to registration, browsing, interactive participation and other behaviors, if the online platform does not invest more costs to operate and protect it after collecting the data, and does not form independent competitive interests, it is not appropriate to give exclusive protection only because the online platform claims rights, otherwise it is easy to cause data closure and data monopoly, which is not conducive to the healthy development of the digital economy.

In a copyright infringement and unfair competition dispute between a computer company and an Internet company,8 A computer company enjoys the copyright and operation rights of the game in question. A network company pre-installed the game software on its 5G cloud gaming platform, and game users can directly log in and run the game on the platform without downloading the game APP. When using the cloud gaming platform, the user agrees and authorizes an online network company to collect user registration information, game data, etc. The Guangzhou Internet Court held that the account information of game users and the relevant data of participating in the game were all original data. In the process of forming these data, the user plays a leading role, and a network company also collects such data after obtaining the user's authorization, without violating the protection measures of a computer company, not hindering the collection and use of game user data by a computer company, or affecting the normal operation of the game involved in the case. Accordingly, the Guangzhou Internet Court found that the collection of raw data by an Internet company did not constitute unfair competition.

2. Determination of improper acquisition and use of data

Article 13, Paragraph 3 of the Anti-Unfair Competition Law (revised in 2025) clearly lists three typical ways of improperly obtaining and using data, including fraud, coercion, avoiding or sabotaging technical management measures, and adding the word "etc" at the end as a cover to deal with the endless emergence of new technical means. Fraud refers to obtaining other people's data through fake websites, phishing emails, deceptive user agreements, and other methods that fabricate facts and conceal the truth. Coercion refers to coercing others to provide data through perpetrators or threats by using dominant positions and technical means. Avoiding or destroying technical management measures refers to bypassing or cracking the technical or management measures of others to capture data.9 The first two behaviors have strong attributability, while the third behavior is more complex, and the determination of this situation and other improper data acquisition behaviors depends on the proof of their imputability. Data scraping that violates platform restrictions may strengthen the platform's control over data if it is considered illegal and does not specifically assess the actual extent of the damage or set clear criteria. This will give the platform the ability to unilaterally determine the permission to open data, thereby hindering data circulation and value release.10

Factors considered in judicial practice include: First, whether the data holder has actually set up technical management measures for the management and protection of data. The second is whether the acquisition behavior is carried out by bypassing technical management measures or by means of adversarial destruction. The third is causation, that is, the behavior actually obtains data against the will of the data holder. Teacher Kong Xiangjun believes that there are many situations of protection measures related to data and information, and there are different protection positions and requirements. The technical measures or management measures in the "Data Clauses" of the Anti-Unfair Competition Act (revised in 2025) are different from those in the Copyright Law, Personal Information Protection Law, and Cybersecurity Law. The technical management measures stipulated in the Data Clauses shall not be limited to protection measures with higher technical standards, but shall also include other management measures, or specifically technical management measures, and shall not exclude non-technical management measures (such as crawler agreements that prohibit scraping). The protection measures of the Data Terms are only necessary to the minimum degree of declaration of rights, and are not required to prevent the acquisition of data information.11

（1）Determination of the impropriety of data scraping behavior

Data scraping behavior is essentially a technical form that uses machines and programs to simulate user access behavior, which can efficiently and cost-effectively filter, filter, and aggregate data required for subsequent use. There are various technical methods of data scraping, including but not limited to: web crawlers, which simulate browser behavior through programs, send HTTP requests to target websites, obtain web page HTML content, and then parse and extract the required data. It is usually suitable for scraping static web content, such as news articles, product information, public data, etc. The principle of API API calls is to use the public APIs provided by websites or platforms to obtain structured data, such as json and xml format data, through HTTP requests. It is usually suitable for platforms with public APIs, such as social media, weather services, product data from e-commerce platforms, etc.; Browser automation tools work by simulating user actions, such as clicking, typing, and scrolling through the browser, to obtain dynamically loaded page data. Typically suitable for complex dynamic web pages.

Common anti-scraping measures include setting robots protocols, dynamic IP blocking, request frequency limiting, user-agent identification and blocking, changing HTML tags from time to time, and password authentication. It can be seen from the above that only the act of avoiding or destroying effective anti-scraping measures or equivalent acts, and thus obtaining and using data against the will of the data holder, meets the behavioral requirements stipulated in the "Data Clauses" and should be deemed to be unjust.

In the case of an unfair competition dispute between an Internet company and an information company,12 A network company is the operator of the social platform involved in the case, and an information company is the operator of the iDataAPI website, and network users can obtain a large amount of data from the social platform involved in the case through the API (application programming interface) provided by the iDataAPI website, which not only covers all the content on the web page of the social platform involved in the case, but also contains a large amount of background service data in the operation and management process of the social platform involved in the case, and stores and sells it. The Guangdong Provincial High People's Court held that an information company fraudulently obtained the call authority of a special interface for transmitting data to the user by disguising the iDataAPI program as a large number of different real Weibo users and using some technical cracking methods to obtain the corresponding data interface address. Accordingly, the court found that the data scraping behavior of an Internet company was improper.

To determine whether the behavior of data capture is legitimate, first examine whether the means of data capture are legitimate. Normally, scraping non-public data with access rights set by operators will inevitably circumvent or destroy the operator's technical management measures, so it is considered an improper scraping behavior. However, if the means, such as the means, of others, are legitimate, it is necessary to make further judgments on the legitimacy of the captured public data based on other factors such as the quantity, scale, and value of the captured data, as well as whether the subsequent use of the captured data has caused damage to the legitimate rights and interests of the operator of the captured data.

In an unfair competition dispute between an Internet company and a technology company,13 The Haidian District Court of Beijing held that although a technology company argued that it had captured the public data in the data involved in the case through legitimate web crawler technology, its scraping behavior was not normally recorded by the server of a network company, proving that it did not use web crawlers or other normal access behaviors to capture the data involved in the case. Combined with the fact that the data involved in the case involves a large amount of data, a technology company should obtain public data by disguising itself as a user login or simulating user behavior by sending requests to the backend server of a network company and parsing it in accordance with browser rules.

In judicial practice, when applying the Anti-Unfair Competition Law to adjust Internet-related competition behavior, the court will not tend to consider static interests and commercial results too much, but based on the legitimacy of the means of competition and the soundness of the competition mechanism, and consider the fundamental goal of market competition.

3. Consequences of damage

Article 13, Paragraph 3 of the Anti-Unfair Competition Law (revised in 2025) stipulates that the elements of damage are to harm the legitimate rights and interests of other business operators and disrupt the order of market competition. harming the legitimate rights and interests of other operators, including seizing other people's traffic and trading opportunities; increase the operating costs of others and reduce the operating capacity; Obstructing others from fulfilling their obligations to maintain network security and protect user privacy; reduce user stickiness and evaluation; derogating goodwill; reduce advertising revenue, membership fee income, etc. Disrupting the order of market competition includes destroying the supply and demand mechanism of the industry, causing an imbalance in the allocation of market resources, increasing market transaction costs, and undermining industry rules. At the same time, unfair competition may also harm the legitimate rights and interests of consumers, such as infringing on the privacy and right to know of network users. Reduced consumer choice; Resulting in a decrease in product or service quality, an increase in price, etc.

In the unfair competition dispute between a technology company and a technology company's "transformation into comic special effects",14The Beijing Intellectual Property Court first analyzed whether a technology company enjoyed the competitive interests of the model of the transformation of comic special effects (hereinafter referred to as the model involved in the case), and held that a technology company invested a lot of operating resources in the research and development of the transformation comic special effects model, and the imaging effect of the transformation comic special effects also underwent a process of continuous optimization. The model (structure and parameters) of a technology company that has achieved innovative advantages, operating income and market benefits and transforms into comic special effects should constitute its competitive interests protected by the Anti-Unfair Competition Law.

Secondly, the sued model used by a technology company has a high degree of identity in terms of technical implementation with the model in question, and the effect of the special effect product produced by the model is extremely similar, and it fails to prove that the accused model and training data come from its own R&D activities. Finally, the court held that the accused model had a strong substitution and diversion effect on the model involved, caused substantial damage to the competitive interests of a technology company, distorted the normal supply and demand mechanism of the transformation into a comic special effects model, disrupted the business activities of the artificial intelligence model and the healthy and orderly competition order, and affected the future choice space of consumers. Accordingly, the court found that the alleged act constituted unfair competition.

To sum up, when applying Article 13, Paragraph 3 of the Anti-Unfair Competition Law (revised in 2025), courts usually follow the logical steps of adjudication of the legality of rights and interests, improper behavior, damage results, and measurement of the interests of all parties when applying Article 13, Paragraph 3 of the Anti-Unfair Competition Law (revised in 2025).

At the same time, data protection legislation emphasizes the circulation and use of data, and control and empowerment are also for better data sharing. Therefore, it is necessary to limit the scope and intensity of legal protection, retain sufficient space for data sharing and public domain, prevent excessive monopoly and control rights, and allow necessary fair use to promote other public policies.15 Mr. Kong Xiangjun believes that although the "Data Clause" does not set exceptions for protection, the exceptions in the application of law include: 1. The amount of data obtained or used is insufficient to cause substantial damage, such as the business activities commonly used in judgments that do not substantially replace the business activities of the right holder; 2. The acquisition or use of the right holder's data does not conflict with the normal business activities of the right holder and does not substantially harm the business interests of the right holder; 3. Other situations where other people's data is obtained or used in a way that does not violate good faith and business ethics. This is to keep a bottom line clause to leave room for the fair use of data in ways such as promoting public interest.16

4. Liability for damages

The determination of liability for damages for unfair competition in data is still determined by the actual loss of the right holder or the benefits obtained by the infringer as a result of the infringement. if it is difficult to determine, the court shall give compensation of up to 5 million yuan. Statutory compensation considerations include data type, number of users, and popularity. the cost of data collection and processing by the right holder; the subjective intention of the infringer, the duration, scope, and impact of the infringement.

In judicial practice, it is difficult to quantify the losses suffered by right holders such as traffic, reduced advertising, reduced user stickiness, and damage to goodwill, and infringers usually do not submit real evidence of infringement profits. Article 95 of the Several Provisions of the Supreme People's Court on Evidence in Civil Litigation stipulates: "If a party refuses to submit evidence of control without justifiable reasons, and the party who bears the burden of proof claims that the content of the evidence is not conducive to the controller, the people's court may find that the claim is established." "The right holder may use the obstruction of evidence rule to provide evidence for the losses suffered by the infringer or the profits of the infringer, such as the appraisal or evaluation report of the loss of advertising and user activity, or the increase in the number of clicks on the infringer's website, the number and amount of products sold, the sales announced during publicity, or the profit margin of operators in the same industry, etc., and the infringer refuses to provide evidence such as accounting books, and may claim to determine the amount of damages based on the evidence submitted by the right holder.

In the above-mentioned unfair competition dispute between an Internet company and an information company,17 An information company announced on the iDataAPI website the total number of calls made by its users to the interface of the social platform involved in the case and the cost per 100 calls, and as of May 29, 2019, the number of calls exceeded 2.1 billion. Based on this, the Shenzhen Intermediate People's Court calculated that the infringement profit income exceeded 20 million yuan. Based on factors such as the many types of unfair competition behavior of an information company, the use of malicious technical means, the long duration, the huge scale of data called, and the serious consequences of damage, it fully supported the 20 million yuan claim for compensation of an online company. The Guangdong Provincial High People's Court held that an information company was fully capable of providing evidence to prove the specific number of calls and revenue of the API interface, including detailed interface call records and charging records, as well as related financial books and materials. However, when an online company clearly filed a claim for compensation of 20 million yuan, the calculation method and related evidence, an information company still chose to only make an oral defense without providing any specific financial evidence to refute it, and still neglected to provide evidence in the second instance. According to the principle of litigation good faith, it is fairer and more reasonable for an information company rather than an online company to bear the corresponding legal consequences of the adverse presumption. The court finally rejected the appeal and upheld the original judgment.