On March 1, the Provisions on the Administration of Algorithm Recommendation of Internet Information Service (hereinafter referred to as the "Provisions"), jointly issued by the Cyberspace Administration of China and four other departments, came into effect, which clearly requires that all kinds of Internet companies providing algorithm recommendation services guarantee users' right to know and choose algorithms, and should provide users with options that do not target their personal characteristics, or conveniently close the options for algorithm recommendation services, and "establish and improve mechanisms for manual intervention and user choice."
As of now, according to incomplete statistics, Apps such as Douyin, Today's Toutiao, WeChat, Taobao, Baidu, Dianping, Weibo, and Xiaohongshu have all launched algorithmic shutdown keys, allowing users to close "personalized recommendations" with one click.
Algorithmic recommendations are convenient, but the shutdown function must be present
Whether "personalized recommendation" is related to it or not is currently becoming a tangle for many people. Some people feel that they see more categories after the closure, and some people think that shopping has lost interest after the closure. Will the algorithm recommend, a once widely used Internet business model, come to an end with the launch of the close button? The news came out, triggering heated discussion among netizens, most of whom agreed with this measure.
Zhang Linghan, associate professor at the University of Science and Technology Beijing, analyzed that before the promulgation and implementation of the "Provisions", e-commerce platforms mostly provided options that did not target individual characteristics by providing price ranking and sales volume ranking. Some head platform companies have also launched the switch options of "turn off targeted advertising push" and "turn off personalized recommendations", but it is often more difficult to find more hidden users. After the implementation of the Personal Information Protection Law, a large number of platform companies have begun to develop and launch such functions.
Zhang Linghan believes that the "Provisions" clearly put forward the "convenience" requirement of closing the algorithm recommendation in order to further facilitate users, and also to make the right given by law to users to reject the algorithm and can "say no" to the implementation. According to Article 24 of the Mainland's Personal Information Protection Law, if it is an automated decision involving the major rights and interests of users, users also have the right to obtain explanations. The purpose of this provision is to lift the "personalized veil" of the information received by the user, so that the user can see the real "organic search results", rather than being in the flow of advertising, news, pricing and other information tailored for themselves. This is essential to protect users' right to know, allowing users to make consumption decisions based on more real and diverse information.
However, the Rule of Law Network Research Institute noted that at present, a considerable number of Apps have also played a little "little clever", and the "personalized recommendation" off button is buried very deeply, which is generally found in privacy and advertising-related setting options. However, Apps such as WeChat and Dianping have made a "personal information collection list" that clearly tells consumers what information is collected and how it is used. Such a move can be called a "conscientious businessman".
In fact, the various personalized recommendations of mobile applications can indeed allow consumers to easily brush up on messages, videos, and products that they are interested in, but they may also unconsciously "contribute" their personal privacy.
Zhang Linghan said that in the past platform practice, not many users chose to turn off personalized push, and many users chose to open it again after closing. According to the survey, users reported that after turning off personalized push, they could not find the goods and information they needed, and the efficiency of searching for information on the Internet was greatly reduced. Because of the "long tail effect", the information placed after the hot spot, if there is no accurate matching of the personalized push algorithm, it is difficult to present the front end of the platform information flow to be seen by the user. On the other hand, such a function as enterprise launching also requires certain financial and human support, and needs to be set in a directory with a higher app level (that is, the user can find this button by opening two or three layers of directories). But the ability to turn off personalization options must exist – this gives the user a choice, a choice that is not "calculated", a choice that is not wrapped up in the flow of information, a choice to obtain information autonomously.
Can turning off "personalized recommendations" curb "big data killing"?
In life, there are often network users who complain that they are "killed by big data". According to the "China Great Security Perception Report (2021)" jointly released by the Internet Development Research Center of Peking University and Internet companies, 70% of respondents feel that algorithms can obtain their own preferences and interests to "calculate" themselves, and nearly 50% of respondents said that they want to escape the network and stay away from mobile phones under the shackles of algorithms.
The Rule of Law Network Research Institute entered the keyword search of "big data killing" on the black cat complaint platform, and the page showed a total of 2317 results. Randomly browse the relevant complaints, including takeaway platforms, e-commerce, online ride-hailing, shared bicycles, travel e-commerce platforms, ticketing platforms, etc.
The Provisions propose that where algorithms recommend service providers to sell goods or provide services to consumers, they shall protect consumers' rights to fair trade, and must not use algorithms to carry out unreasonable differential treatment and other illegal acts on transaction conditions such as transaction prices based on consumers' preferences and trading habits.
So, can turning off "personalized recommendations" curb "big data killing"?
In this regard, Zhang Linghan analyzed, first of all, the use of algorithms to make unreasonable and unfair decisions, harming the interests of consumers is prohibited by law. Big data is both sides of the same coin. In mainland laws and regulations, in order to attract new customers, for promotional purposes, etc., it is possible to differentiate pricing. However, if the e-commerce platform uses the data advantage and technical advantages of the user to cause the user to suffer unfair and unreasonable pricing, it is regulated by law. For example, in a case at the Keqiao Court in Shaoxing, Zhejiang Province, last year, in which a regular customer of a platform enterprise was subjected to double pricing under the promise of the "lowest price agreement", the plaintiff's claim, that is, the user's litigation claim was supported.
Secondly, curbing big data killing and closing personalized recommendations is only one link in systematic governance. The effect of turning off personalized recommendations is to "cut off" the algorithm, that is, not to provide their own data to the algorithm for differentiated pricing, and there are still many other data sources and methods to achieve "big data killing". Therefore, to root out the "big data killing", the key is to unveil the veil of technology neutrality and examine the subjective intentions of algorithm recommendation service providers.
In addition, it is also necessary to build an algorithmic audit system so that users have a more effective way to prove that they have been killed by big data, and also make the administrative law enforcement of the regulatory authorities have a basis.
Experts: Standardized algorithm recommendation chaos, the need to land detailed rules
For how to increase efforts to regulate the relevant APP algorithm recommendation chaos in the future, Zhang Xin, executive director of the Digital Economy and Legal Innovation Research Center of the University of International Business and Economics, believes that first of all, the detailed rules for landing should be further promoted as soon as possible in accordance with the "Provisions". At present, for Internet enterprises, the algorithm compliance framework is only preliminarily established, not only lacking corresponding reference samples, but also confusing for the understanding of some principled laws and regulations. Therefore, it is recommended that stakeholders such as regulators, industry organizations and platform enterprises work together to promote the detailed rules that need to be implemented in the provisions on the management of algorithm recommendation, and try to give enterprises clearer compliance guidelines.
Second, the effective communication between business departments and compliance departments should be strengthened within the enterprise, and an algorithmic compliance and risk management framework should be established as soon as possible. Promote the compliance practice of algorithm recommendation as soon as possible from the aspects of enterprise privacy policy, APP function design, algorithm mechanism mechanism review, etc.
Finally, it is recommended that the majority of users also actively exercise the right to a new type of algorithm. For example, to obtain explanations and explanations for algorithm recommendation, app that has not yet provided algorithm recommendation shutdown function, and provide convenient and easy algorithm recommendation shutdown function as soon as possible.
Source: Legal Network