The Cyberspace Administration regulates algorithm recommendations and prohibits the implementation of traffic fraud control hot search
2021-08-28: [Original Article Link].
Jiang Dan, Securities Times
The 'information cocoon' and 'emotional contagion' effect brought by algorithm recommendation is expected to be cracked.
On August 27, the State Internet Information Office on the 'Internet information services algorithm recommendation management regulations (draft for public comment)' for public comment, requiring the algorithm recommendation service providers shall not use the algorithm recommendation service to engage in activities prohibited by laws and administrative regulations, such as endangering national security, disrupting the economic and social order, infringing on the legitimate rights and interests of others, and shall provide users with options that do not target their personal characteristics, or provide users with a convenient option to turn off the algorithm recommendation service. The user is provided with the option to conveniently turn off the algorithm recommendation service. The deadline for feedback is September 26, 2021.
Algorithmic recommendations have become the 'basic operation' of each platform. According to the draft, the regulations will apply to the application of algorithmic recommendation technology to provide Internet information services in the territory of the company.
The draft proposed to encourage relevant industry organizations to strengthen industry self-regulation, establish and improve the self-regulatory system and industry guidelines, organize the development of industry standards, and supervise and guide algorithmic recommendation service providers to establish and improve service specifications, provide services in accordance with the law and accept social supervision.
For the algorithm recommendation easily lead to youth addiction and blind imitation of the problem. The draft of the consultation points out that the algorithm recommendation service providers to provide services to minors, should be in accordance with the law to fulfill the network protection obligations of minors, and by developing a model suitable for minors to use, provide services suitable for the characteristics of minors and other ways to facilitate access to information content for the physical and mental health of minors.
Algorithm recommendation service providers shall not push information content that may trigger minors to imitate unsafe behavior, as well as violate social morality, induce bad habits of minors, etc., which may affect the physical and mental health of minors, and shall not use algorithm recommendation services to induce minors to become addicted to the Internet.
The draft also clarifies that the algorithm recommendation service providers shall not use the algorithm to falsely register accounts, illegal trading accounts, manipulation of user accounts, or false likes, comments, retweets, web navigation, etc., the implementation of traffic falsification, traffic hijacking; shall not use the algorithm to block information, excessive recommendations, manipulation of the list or search results sorting, control of hot search or selected and other intervention information presentation, the implementation of self-preferential treatment, unfair competition, influence network public opinion or circumvent regulation.
In early August, the central propaganda department and other five departments jointly issued the 'guidance on strengthening the work of the new era of literature and art criticism' has been clear, strengthen the research and guidance of network algorithms, network algorithm recommendations to carry out comprehensive governance, do not provide channels for the dissemination of wrong content.
For 'big data to kill ripe', this document again stressed that the algorithm recommended service providers to sell goods or provide services to consumers, should protect the legitimate rights and interests of consumers, shall not be based on consumer preferences, trading habits and other characteristics, the use of algorithms in the transaction price and other trading conditions to implement unreasonable differential treatment and other illegal acts.
Not only that, the draft for comments on algorithmic recommendations to the user set discriminatory or prejudicial labeling behavior is prohibited. According to the regulations, the algorithm recommendation service providers should strengthen the management of user models and user tags, improve the rules of interest points credited to the user model, not to put illegal and undesirable information keywords into user interest points or as user tags and push information content accordingly, and not to set discriminatory or biased user tags.
The recently voted personal information protection law requires that information pushing and commercial marketing to individuals through automated decision-making should be accompanied by options that do not target their personal characteristics or provide individuals with a convenient way to reject them.
This is reflected in the consultation draft, which requires that algorithmic recommendation service providers should provide users with the option not to target their personal characteristics, or provide users with a convenient option to turn off the algorithmic recommendation service. Users choose to close the algorithm recommendation service, the algorithm recommendation service providers should immediately stop providing related services.
According to the draft, the state Internet information department to establish a classification and grading management system, according to the algorithm recommendation service of public opinion attributes or social mobilization capacity, content categories, user size, algorithm recommendation technology to deal with the degree of data sensitivity, the degree of intervention in user behavior and other algorithm recommendation service providers to implement classification and grading management.
Xue Jun, director of E-commerce Law Research Center of Peking University, believes that, in addition to clear legal provisions, some kind of socialized review mechanism can be established to evaluate the consequences of the platform's use of algorithms and require the platform to be optimized based on public values.
Note: This is auto-translated version of an article meant for Chinese audience. A mature and nuanced reading is suggested.