Provisions on the Administration of Internet Information Service Algorithmic Recommendation (Draft for Solicitation of Comments)

Article 1: These Provisions are drafted on the basis of the Cybersecurity Law of the PRC, the PRC Data Security Law, the Personal Information Protection Law of the PRC, The Measures on the Administration of Internet Information Services, and other relevant laws and administrative regulations, so as to regulate internet information services algorithmic recommendation activities, preserve national security and the societal public interest, protect the lawful rights and interests of citizens, legal persons, and other organizations, promote the healthy development of internet information services, and carry forward the Core Socialist Values.

Article 2: These provisions apply to the use of algorithmic recommendation technology to provide internet information services within the mainland territory of the People's Republic of China. Where laws or administrative regulations otherwise provide, follow those provisions.

"Use of algorithmic reccomendation technology" as used in the preceding paragraph refers to the use of algorithm technologies types such as generation and synthesis, individualized pushing, sequence refinement, search filtering, schedule decision-making, and so forth to provide users with information content.

Article 3: The State Internet Information Office is responsible for oversight, management, and law enforcement efforts on algorithmic recommendation services nationwide. Based on their duties, provincial, autonomous region, and directly governed municipalities' internet information departments are responsible for oversight, management, and law enforcement efforts on algorithmic recommendation services in the corresponding administrative region.

Article 4: Algorithmic recommendation service providers providing algorithmic recommendation services shall obey laws and regulations, respect social mores and ethics, obey commercial and professional ethics, follow the principles of equity and fairness, openness and transparency, being rational and reasonable, and good faith.

Article 5: Relevant industry organizations are encouraged to strengthen industry discipline, establishing and completing systems for self-discipline and industry conduct codes, organizing the drafting of industry standards, promoting and guiding algorithmic recommendation service providers to establish and complete service specifications, to provide services in accordance with law, and accept societal oversight.

Article 6: Algorithmic recommendation service providers shall persist in being oriented towards mainstream values, optimize mechanisms for algorithmic recommendation services, actively transmit positive energy, and promote the uplifting use of algorithms.

Algorithmic recommendation service providers shall implement entity responsibility for algorithms; establish and complete management systems such as for user registration, reviews of information for publication, reviews of algorithm mechanisms, security assessment and monitoring, security incident response and handling, data security protections, and protection of personal information; draft and disclose service rules for algorithmic recommendations; and allot professional personal and technical support corresponding the scal of algorithmic recommendation services.

Article 7: Algorithmic recommendation service providers shall implement entity responsibility for algorithms; establish and complete management systems such as for user registration, reviews of information for publication, reviews of algorithm mechanisms, security assessment and monitoring, security incident response and handling, data security protections, and protection of personal information; draft and disclose service rules for algorithmic recommendations; and allot professional personal and technical support corresponding to the scale of algorithmic recommendation services.

Article 8: Algorithmic recommendation service providers shall periodically review, assess, and verify algorithm mechanisms, models, data, and outcomes, and must not set up algorithmic models that induce users to become addicted or spend large amounts, or that go against good public customs.

Article 9: algorithmic recommendation service providers shall strengthen the management of information content, establishing and completing a pool of characteristics used to identify illegal and negative content, and improve standards, rules, and procedures for inclusion in the pool. Where it is discovered that algorithmically generated or synthesized information is not conspicuously labeled, transmission may continue on after making a conspicuous label.

Where unlawful information is discovered, transmission shall be immediately stopped, and measures shall be employed to eliminate it or otherwise address it, stop the spread of the information, store relevant records, and report to the internet information departments. Where negative information is discovered, it shall be addressed in accordance with the relevant provisions on the management of the internet information content ecology.

Article 10: Algorithmic recommendation service providers shall strengthen the management of user modeling and user labels, improving rules for recording the interests of user models; keywords of unlawful and negative information must not be used as interests or user labels used to push information content, and discriminatory or stereotyping user labels must not be established.

Article 11: Algorithm recommendation service providers shall strengthen the page ecological management of algorithm recommendation service pages, establish and improve mechanisms for manual intervention and users' independent selections, and actively present information content that comports with mainstream value orientations ​​in key steps such as the homepage, home screen, hot searches, selections, top content lists, and pop-up windows.

Article 12: Algorithmic recommendation service providers shall comprehensively use strategies such as for eliminating duplicate content and for fragmentation and intervention, and optimize the transparency and explainability of rules for searches, sorting, selections, pushing, and displays, to avoid producing a negative impact on consumers causing contention and disputes.

Article 13: algorithmic recommendation service providers must not use algorithms to register fake accounts, illegal trade accounts, manipulate user accounts, or give false likes, comments, forwards, web page navigation, and so forth, or to carry out traffic fraud or traffic hijacking; and must not use algorithms to interfere with information presentation such as by blocking information, making excessive recommendations, manipulating the order of top content lists or search results, or controlling hot searches or selections, to carry out self-preferential treatment, unfair competition, influence online public opinion, or evade oversight.

Article 14: Algorithmic recommendation service providers shall inform users in a conspicuous fashion of the circumstances of the algorithmic recommendation services provided, and display the algorithmic recommendation services' basic principles, purposes, operation mechanisms, and so forth in an appropriate manner.

Article 15: The providers of algorithmic recommendation services shall provide users with options not targetting their individual characteristics or provide users with convenient options to close algorithmic recommendation services. Where users select to close algorithmic recommendation services, the algorithmic recommendation service providers shall immediately stop providing those services.

Algorithmic recommendation service providers shall provide users with functions for selecting, revising, or deleting user labels used in algorithmic recommendation services.

Where users find that algorithmic recommendation service providers' use of algorithms has caused a major impact on their rights and interests, they have the right to request that the algorithmic recommendation service providers give an explanation and employ corresponding corrective or remedial measures.

Article 16: Where algorithmic recommendation service providers provide services to minors, they shall perform obligations to protect minors online in accordance with law, and facilitate minors obtaining beneficial and health information content through methods such as having modes appropriate for use by minors and providing services suited to the characteristics of minors.

Algorithmic recommendation service providers must not push information content to minors that might impact minors' physical and psychological health such as possibly leading them to imitate unsafe behaviors and conduct contrary to social mores or inducing negative habits, and must not use algorithmic recommendation services to induce minors' addiction to the internet.

Article 17: Algorithmic recommendation service providers providing work scheduling services to laborers shall establish and improve algorithms for platform order allocation, the composition and payment of compensation, work times, awards and punishments, and so forth, to perform obligations to protect laborers' rights and interests.

Article 18: Where algorithmic recommendation service providers market goods or provide services to consumers, they shall protect the consumers' lawful rights and interests and must not use algorithms to carry out unreasonable differentiation in treatment in terms of transaction prices or other transaction conditions or carry out other unlawful conduct, based on consumers preferences, transaction habits, or other traits.

Article 19: the state internet information department is to establish a categorical and hierarchical management system to conduct management by category and grade of algorithmic recommendation service providers based on the algorithmic recommendation services' public opinion properties or capacity for social mobilization, the content types, the scale of users, the sensitivity of the data handled by the algorithmic recommendation technology, the degree of interference in user conduct, and so forth.

Article 20: Within 10 working days of beginning to provide services, algorithmic recommendation service providers that have public opinion properties or capacity for social mobilization shall provide the service provider's name, service form, field of application, type of algorithm, algorithm self-assessment report, content to be displayed, and other such information through the internet information services algorithm filing system, and perform filing procedures.

When algorithmic recommendation service providers' filed information changes, they shall complete modification procedures within 5 working days of the change.

Where algorithmic recommendation service providers terminate services, they shall handle procedures to deregister the filings within 3 working days of the termination, and make appropriate arrangements.

Article 21: After the state and provincial, autonomous region, or directly governed municipality internet information departments receive the materials for filing from a filer, where the materials are complete, they shall file them within 30 working days and issue a file index number and display them; where the materials are incomplete, they are not to file them, and shall notify the filer and explain the reasons within 30 working days.

Article 22: algorithmic recommendation service providers that have completed filing shall indicate their file index number and provide links to displayed information in a conspicuous location in their websites or applications that provide external services.

Article 23: Algorithmic recommendation service providers that have public opinion properties or capacity for social mobilization shall carry out security assessments in accordance with relevant state provisions.

Algorithmic recommendation service providers shall improve mechanisms for the management of algorithmic recommendation services, retaining information such as algorithmic recommendation service logs for a retention period of at least 6 months, and lawfully providing them to relevant law enforcement departments that make inquiries.

Article 24: The state and provincial, autonomous region, or directly governed municipality internet information departments, in conjunction with relevant regulatory departmnets, are to carry out algorithm security assessments and oversight inspections of algorithmic recommendation services, and promptly submitting comments and time limits on corrections of any issues discovered.

Algorithmic recommendation service providers shall cooperate with the relevant competent departments conducting security assessments and oversight inspections in accordance with law, and provide necessary technological and data support and assistance.

Article 25: Relevant bodies and personnel participating in security assessments and oversight inspections shall strictly keep the confidentiality of personal information, private information, and commercial secrets which they learn of in performing their duties, and must not leak, sell, or unlawfully provide it to others.

Article 26: algorithmic recommendation service providers shall accept social oversight, set up convenient portals for complaints and reports, and promptly handle the public's complaints and reports.

Algorithmic recommendation service providers shall establish channels and systems for user appeals, to standardize the handling of user appeals and give prompt feedback, to truly safeguard users' lawful rights and interests.

Article 27: Where algorithmic recommendation service providers violate articles 7, 8, paragraph 1 of article 9, articles 10, 11, 12, 13, 14, the second paragraph of article 15, article 22, or article 26 of these Provisions, the state and provincial, autonomous region, or directly governed municipality internet information departments are to give warnings, circulate criticism, and order corrections in a set period of time on the basis of their duties; where the corrections are refused or the circumstances are serious, they are to be ordered to suspend information updates and given a concurrent fine of between 50,000 and 300,000 RMB. Where violations of public security are constituted, a public security administrative sanction is lawfully given; where a crime is constituted, criminal responsibility is to be pursued in accordance with law.

Article 28: Where algorithmic recommendation service providers violate article 6, paragraph 2 of article 9, paragraphs 1 or 3 of article 15, articles 16, 17, 18, 23, or paragraph 2 of article 24 of these Provisions, the competent departments such as for internet information are to address it on the basis of their duties and in accordance with relevant laws, administrative regulations, and department rules.

Article 29: Where algorithmic recommendation service providers with public opinion properties or capacity for social mobilization violate article 20 of these Provisions by failing to file as requested or by concealing relevant circumstances, providing false materials when reporting for filing, or by obtaining filing through improper tactics such as fraud or bribery, the state or provincial, autonomous region, or directly governed municipality internet information departments are to revoke the filings in accordance with law and give warnings, circulate criticism, and order corrections to be made in a set period of time; where corrections are refused or the circumstances are serious, they are to order a suspension of information updates and give a concurrent fine of between 5,000 and 30,000 RMB.

Where algorithmic recommendation service providers with public opinion properties or capacity for social mobilization that terminate services fail to promptly handle procedures for deregistering filings, or where they receive administrative punishments for serious violations such as having their internet information service permits cancelled, websites shut down, or services terminated, the state or provincial, autonomous region, or directly governed municipality internet information departments are to deregister their filings.

Article 30: These Provisions shall come into effect on XX, XX, 2021.