Summary
Overview & Analysis
Source of the "public opinion / social mobilization" filing trigger: This document is the origin of the algorithm filing requirement that is referenced by the Generative AI Services Measures, the Deep Synthesis Provisions, the AI Content Labeling Measures, and the CII Security Regulation. Any provider with public opinion attributes or social mobilization capabilities must complete algorithm filing under this instrument — and all subsequent AI regulations in this series that require "algorithm filing" point here.
The Provisions on the Administration of Internet Information Service Algorithmic Recommendation are China's foundational regulation for the governance of algorithmic recommendation services — covering the use of generative synthesis, personalized push, ranking and selection, retrieval and filtering, and scheduling and decision-making algorithms to provide information to users. The Provisions impose obligations across three areas: information service standards (requiring mainstream value orientation, algorithm security responsibility, no addiction-inducing models, no fake accounts or ranking manipulation); user rights protection (opt-out, non-personalized options, tag deletion, explanations for significant impacts, protections for minors, elderly users, gig workers, and consumers); and a filing and security assessment regime for providers with public opinion attributes or social mobilization capabilities. The Provisions also prohibit using algorithms to engage in unfair competition, monopoly conduct, or interference with other providers' lawful operations.
This regulation is particularly relevant for AI because AI recommendation systems are the backbone of virtually every consumer-facing digital product in China — from news feeds and short video platforms to e-commerce search, social media, online education, and customer service chat. For AI project managers and product teams at multinational companies, compliance requires not only understanding the filing and assessment process but also designing algorithm transparency, user control, content moderation, and ethical model governance into products from the outset. The Provisions are explicitly cross-referenced in multiple other instruments in this series — and any AI product that generates, filters, ranks, or recommends content to users in China falls within their scope.
Five Covered Algorithm Technology Types (Article 2)
Type 1
Generative Synthesis
GenAI content, text & image generation
Type 2
Personalized Push
News feeds, social content, targeted advertising
Type 3
Ranking & Selection
Trending topics, featured content, editorial picks
Type 4
Retrieval & Filtering
Search results, content moderation, spam filtering
Type 5
Scheduling & Decision-Making
Gig work dispatch, pricing algorithms, supply chain
Algorithm Filing Timeline (Articles 24–26)
Before / At Service Launch
Within 10 Working Days
File via algorithm filing system: name, service form, application field, algorithm type, self-assessment report, proposed disclosure
On Change to Filing Info
Within 10 Working Days
Update filing information within 10 working days of any change
Authority Review
Within 30 Working Days
CAC completes filing, issues number, and publicly discloses; or rejects with reasons
On Service Termination
Within 20 Working Days
Deregister filing and make proper arrangements for users
Key Prohibitions — Article 14 Algorithm Manipulation Rules
- Creating false accounts, illegally trading accounts, or manipulating user accounts using algorithms
- Fabricating likes, comments, or reposts using algorithms
- Using algorithms to suppress information or push content in excess of normal recommendation
- Manipulating rankings, search result order, trending topics, or featured content to influence public opinion
- Using algorithms to evade regulatory supervision or interfere with information presentation
- Using algorithms to impose unreasonable restrictions on other internet information service providers (Art. 15)
- Using algorithms to engage in monopolistic or unfair competition practices (Art. 15)
Four Core User Rights (Chapter III)
Article 16
Prominent disclosure of algorithmic recommendation use, basic principles, purposes, and main operating mechanisms
Transparency right
Article 17
Non-personalized options OR convenient opt-out; immediate cessation when user chooses to disable
Choice & opt-out right
Article 17
Select or delete user tags used for algorithmic recommendation based on personal characteristics
Tag deletion right
Article 17
Explanation and assumption of corresponding responsibilities where algorithm use has a significant impact on user rights and interests
Explanation right
Relevant AI Scenarios
These Provisions apply to all AI scenarios involving algorithm recommendation technology — from large-scale consumer platforms to enterprise AI systems that rank, filter, select, or push content or decisions to users. Any AI product that personalizes content, sorts search results, optimizes work dispatch, controls trending content, or generates and recommends information in China is in scope.
1. News, Advertising & Social Platform Recommendation Services
AI algorithm recommendations are widely used in news feeds, personalized advertising, and social content curation. These platforms must ensure that recommended content aligns with mainstream values, that genuinely unlabeled AI-generated synthetic content is labeled before further transmission, and that users have clear opt-out and non-personalized recommendation options. Providers with public opinion attributes or social mobilization capabilities must complete algorithm filing before offering services.
2. E-Commerce Search, Pricing & Product Recommendation
On cross-border and domestic e-commerce platforms, AI algorithms recommend products and services based on user interest, transaction history, and behavioral profiling. Article 21 explicitly prohibits using algorithms to implement unreasonable differential treatment in transaction conditions — including price discrimination — based on user preferences or habits. Platforms must provide transparent opt-out mechanisms and ensure pricing algorithms do not create structural discrimination against user segments.
3. Generative AI and Content Synthesis Services
Article 2 expressly includes generative synthesis algorithms within the scope of these Provisions. Article 9 requires providers to establish databases for identifying illegal and harmful content, and where algorithmically generated synthetic content is discovered without conspicuous labeling, it must be clearly labeled before further transmission. Article 13 prohibits generating or synthesizing false news. This makes the Provisions foundational to the AI content labeling regime established by the Deep Synthesis Provisions and the 2025 AI Content Labeling Measures.
4. Gig Economy, Work Dispatch & Algorithmic HR Management
Article 20 specifically addresses providers offering work-dispatch services to workers, requiring protection of workers' lawful rights to remuneration and rest, and improvement of algorithms governing order allocation, compensation structure and payment, working hours, and rewards and penalties. This is directly relevant to platforms using AI for driver dispatch, delivery routing, freelance task allocation, HR scheduling, and performance assessment — making these Provisions among the broadest in the series in terms of the workplace AI scenarios they reach.
5. Education Platforms, Youth AI Services & Elderly-User AI
Article 18 imposes specific protections for minors — requiring age-appropriate modes, prohibiting content that induces unsafe behavior, and prohibiting algorithm-driven addiction to online services. Article 19 requires elderly-user-friendly intelligent services and monitoring for telecom fraud targeting older users. Education platforms, children's content services, and AI tools targeting elderly populations must implement dedicated content controls, age-appropriate recommendation architectures, and fraud-detection measures specifically designed for these user groups.
Practical Advice for Managers at Multinational Companies
These Provisions provide a clear compliance framework for AI recommendation systems in China. Ensuring content transparency, compliant data management, and social responsibility not only reduces legal risk but also enhances platform credibility and user trust — driving the healthy development of AI technologies in the Chinese market.
01
Determine Filing Obligation Before Any Consumer-Facing AI Launch
The filing requirement applies to providers with "public opinion attributes or social mobilization capabilities" — a deliberately broad standard. Any AI product that recommends content to a large user base in China, shapes what information users see, or has the potential to influence public discourse should be assessed for this trigger before launch. Filing must be completed within 10 working days of providing services. Launching without filing — or updating filing late when the product changes — creates direct legal exposure and is one of the most common compliance gaps multinational AI teams face.
02
Build User Control Features into Algorithm-Driven Products from Day One
Articles 17–22 require: prominent disclosure of algorithmic recommendation use; options to receive non-personalized recommendations or turn the service off entirely; ability to select or delete personal tags; explanation mechanisms for significant algorithmic impacts; minors' mode; elderly-friendly mode; and worker rights in dispatch algorithms. These cannot be retrofit after launch — they require product architecture decisions. Require product managers to complete an "algorithm rights checklist" at the design stage for every China-facing AI feature.
03
Establish Routine Algorithm Auditing and Ethics Review Processes
Article 8 requires providers to regularly review, evaluate, and verify algorithm mechanisms, models, data, and application outcomes, and prohibits algorithm models that induce addiction or excessive consumption in violation of law or ethics. Operationalize this as: a defined review cadence for major recommendation models (at least quarterly for consumer-facing products); clear internal criteria for what constitutes an addiction-inducing or excessive-consumption-inducing design; and a documented ethics review process that signs off on algorithm changes. Audits must cover not just technical outputs but the behavioral effects of the algorithm on user patterns.
04
Treat Article 14's Prohibition List as a Non-Negotiable Design Constraint
The prohibited behaviors in Article 14 — fake account creation, fabricated engagement, suppression of information, manipulation of rankings or trending topics, evading supervision — are absolute prohibitions, not a "best efforts" standard. For multinational platforms that use AI to manage content distribution, search relevance, and trending content at scale, every algorithmic feature that touches ranking, trending, or content surfacing must be reviewed against this list. This includes features that may seem commercially neutral, such as "personalized trending" or "quality-boosting" signals that could functionally suppress or amplify content in ways that fall within Article 14's scope.
These Provisions are foundational — not only in their own right, but as the instrument referenced by almost every subsequent China AI regulation in this series. They establish the filing trigger, the user transparency requirements, the algorithm security responsibility framework, the prohibition on addiction-inducing model design, and the user opt-out rights that are built upon and extended by the Deep Synthesis Provisions, the Generative AI Services Measures, the AI Content Labeling Measures, and the Network Data Security Regulation. Any company operating AI-powered services in China that has not mapped its products against these Provisions should treat that review as a baseline compliance step, not an advanced one.
Full Text
Complete Regulatory Text
Promulgated December 31, 2021 · Effective March 1, 2022 · Order No. 9 · Source: Cyberspace Administration of China
Jointly Issued By (4 Authorities)
Cyberspace Administration of China (CAC) · Ministry of Industry and Information Technology (MIIT) · Ministry of Public Security · State Administration for Market Regulation (SAMR)
Table of Contents
- Chapter I — General Provisions (Articles 1–5)
- Purpose, scope & five algorithm types, regulatory framework, core principles, industry self-discipline
- Chapter II — Information Service Standards (Articles 6–15)
- Mainstream value orientation, algorithm security responsibility, regular audit requirement, labeling of unlabeled synthetic content, user model management, key-position editorial duties, Article 14 prohibitions, anti-competition rules
- Chapter III — Protection of User Rights and Interests (Articles 16–22)
- Transparency, opt-out & non-personalized options, tag deletion, minors, elderly, workers, consumers, complaints
- Chapter IV — Supervision and Administration (Articles 23–30)
- Graded & classified system, filing timeline, security assessment, inspections, confidentiality, complaints
- Chapter V — Legal Liability (Articles 31–33)
- Chapter VI — Supplementary Provisions (Articles 34–35)
Chapter I — General Provisions
Article 1 — Purpose and Legal Basis
In order to regulate algorithmic recommendation activities in internet information services, promote core socialist values, safeguard national security and public interests, protect the lawful rights and interests of citizens, legal persons, and other organizations, and promote the healthy and orderly development of internet information services, these Provisions are formulated in accordance with the Cybersecurity Law, the Data Security Law, the Personal Information Protection Law, the Administrative Measures for Internet Information Services, and other laws and administrative regulations.
Article 2 — Scope and Definition of Algorithm Technologies
These Provisions shall apply to the provision of internet information services through the application of algorithmic recommendation technology within the territory of the People's Republic of China (hereinafter referred to as "algorithmic recommendation services"). Where laws and administrative regulations provide otherwise, such provisions shall prevail.
"Application of algorithmic recommendation technology" refers to the use of algorithmic technologies such as generative synthesis, personalized push, ranking and selection, retrieval and filtering, and scheduling and decision-making to provide information to users.
Article 3 — Regulatory Framework
The national cyberspace administration department shall be responsible for the overall coordination of governance and supervision of algorithmic recommendation services nationwide. Relevant departments of the State Council, including those in telecommunications, public security, and market regulation, shall be responsible for supervision and administration within their respective duties.
Local cyberspace administration departments shall be responsible for overall coordination within their respective administrative regions.
Article 4 — Core Principles
Providers of algorithmic recommendation services shall comply with laws and regulations, respect social morality and ethics, observe business ethics and professional ethics, and follow the principles of fairness, impartiality, openness and transparency, scientific rationality, and good faith.
Article 5 — Industry Self-Discipline
Relevant industry organizations are encouraged to strengthen industry self-discipline, establish and improve industry standards, codes of conduct, and self-regulatory systems, and urge and guide providers to formulate and improve service standards, provide services in accordance with the law, and accept social supervision.
Chapter II — Information Service Standards
Article 6 — Mainstream Value Orientation
Algorithmic recommendation service providers shall adhere to mainstream value orientation, optimize service mechanisms, actively disseminate positive energy, and promote the healthy and beneficial application of algorithms.
Providers shall not use such services to engage in activities prohibited by laws and administrative regulations that endanger national security and public interests, disrupt economic and social order, or infringe upon the lawful rights and interests of others; nor shall they disseminate prohibited information, and they shall adopt measures to prevent and resist the dissemination of undesirable information.
Article 7 — Primary Responsibility for Algorithm Security
Algorithmic recommendation service providers shall implement primary responsibility for algorithm security, establish and improve management systems and technical measures for: algorithm mechanism review; technology ethics review; user registration; information publication review; data security and personal information protection; prevention of telecom and online fraud; security assessment and monitoring; and emergency response to security incidents. They shall formulate and publicly disclose relevant service rules, and be equipped with professional personnel and technical support commensurate with the scale of their services.
Article 8 — Regular Algorithm Audit and No Addiction-Inducing Models
Algorithmic recommendation service providers shall regularly review, evaluate, and verify algorithm mechanisms, models, data, and application outcomes, and shall not set algorithm models that induce user addiction or excessive consumption in violation of laws, regulations, or ethical norms.
Article 9 — Information Security: Labeling Unlabeled Synthetic Content
Providers shall strengthen information security management and establish feature databases for identifying illegal and harmful content. Where algorithmically generated synthetic information is found without conspicuous labeling, it shall be clearly labeled before further transmission.
Where illegal information is discovered, transmission shall be immediately stopped, disposal measures taken, records preserved, and reports made. Where harmful information is discovered, it shall be handled in accordance with provisions on online information content ecosystem governance.
Article 10 — User Model and Tag Management
Providers shall strengthen management of user models and user tags, improve rules governing interest points in user models and user tag management, and shall not include illegal or harmful information keywords as user interests or tags for the purpose of content recommendation.
Article 11 — Key Position Editorial Duties
Providers shall strengthen ecological management of service interfaces and pages, establish and improve mechanisms for manual intervention and user choice, and actively present information aligned with mainstream value orientation in key positions such as homepage first screen, trending searches, featured content, rankings, and pop-ups.
Article 12 — Transparency and Content Diversification
Providers are encouraged to comprehensively apply strategies such as content deduplication and diversification, and optimize the transparency and explainability of rules for retrieval, ranking, selection, recommendation, and display, to avoid adverse impacts on users and prevent and reduce disputes.
Article 13 — Internet News Information Services
Where providers offer internet news information services, they shall obtain the relevant licenses, standardize news activities, shall not generate or synthesize false news information, and shall not disseminate news information from entities not authorized by the State.
Article 14 — Prohibited Algorithm Manipulation Behaviors
Providers shall not use algorithms to:
(1) Create false accounts, trade accounts illegally, or manipulate user accounts;
(2) Fabricate likes, comments, or reposts;
(3) Suppress information or excessively recommend content;
(4) Manipulate rankings or search result order, control trending topics or featured content, or otherwise interfere with information presentation;
(5) Implement behaviors influencing online public opinion or evading supervision and management.
Article 15 — Anti-Monopoly and Fair Competition
Providers shall not use algorithms to impose unreasonable restrictions on other internet information service providers, or hinder or disrupt their lawful operations, or engage in monopolistic or unfair competition practices.
Chapter III — Protection of User Rights and Interests
Article 16 — Transparency Disclosure Obligation
Providers shall prominently inform users of the use of algorithmic recommendation services and, in an appropriate manner, disclose the basic principles, purpose, and main operating mechanisms of such services.
Article 17 — Opt-Out, Non-Personalized Options, Tag Deletion, and Explanation Rights
Providers shall provide users with options that are not based on their personal characteristics, or convenient options to turn off algorithmic recommendation services. Where users choose to disable such services, providers shall immediately cease providing them.
Providers shall also provide users with the ability to select or delete user tags used for algorithmic recommendation based on personal characteristics.
Where the use of algorithms has a significant impact on users' rights and interests, providers shall provide explanations in accordance with the law and assume corresponding responsibilities.
Article 18 — Protection of Minors
Providers offering services to minors shall fulfill obligations for the protection of minors online, provide modes and services suitable for minors, and facilitate access to information beneficial to their physical and mental health.
Providers shall not push content that may induce minors to imitate unsafe behavior, violate social morality, or develop harmful habits, nor shall they induce minors to become addicted to online services.
Article 19 — Protection of Elderly Users
Providers offering services to elderly users shall protect their lawful rights, fully consider their needs in travel, healthcare, consumption, and public services, provide age-friendly intelligent services in accordance with national regulations, and carry out monitoring, identification, and handling of telecom fraud information to facilitate safe use of services by elderly users.
Article 20 — Protection of Workers in Dispatch Services
Providers offering work-dispatch services to workers shall protect workers' lawful rights such as remuneration and rest, and establish and improve algorithms related to order allocation, compensation structure and payment, working hours, and rewards and penalties.
Article 21 — Protection of Consumer Fair Transaction Rights
Providers selling goods or services to consumers shall protect consumers' rights to fair transactions, and shall not use algorithms to implement unreasonable differential treatment in transaction conditions such as pricing based on user preferences or transaction habits.
Article 22 — Complaint and Reporting Channels
Providers shall establish convenient and effective channels for user appeals, public complaints, and reporting, clearly define processing procedures and response timelines, and promptly accept, handle, and provide feedback on results.
Chapter IV — Supervision and Administration
Article 23 — Graded and Classified Security Management System
The cyberspace administration department, together with telecommunications, public security, and market regulation departments, shall establish a graded and classified security management system for algorithms, implementing differentiated management based on: public opinion attributes, social mobilization capability, content category, user scale, importance of processed data, and degree of behavioral intervention.
Article 24 — Algorithm Filing Requirements (Public Opinion / Social Mobilization Trigger)
Providers with public opinion attributes or social mobilization capabilities shall, within ten working days of providing services, complete filing procedures through the algorithm filing system by submitting: provider name, service form, application field, algorithm type, algorithm self-assessment report, and proposed disclosure content.
Changes to filing information shall be updated within ten working days of the change.
Termination of services shall be followed by deregistration within twenty working days, with proper arrangements made.
Article 25 — Filing Processing Timeline
Where filing materials are complete, authorities shall complete filing within thirty working days, issue a filing number, and publicly disclose it. Where materials are incomplete, filing shall not be accepted, and reasons shall be provided within thirty working days.
Article 26 — Display of Filing Number
Filed providers shall prominently display the filing number and a disclosure link on their websites or applications.
Article 27 — Security Assessment Obligation
Providers with public opinion attributes or social mobilization capabilities shall conduct security assessments in accordance with relevant national provisions.
Article 28 — Supervision Inspections and Log Retention
Authorities shall conduct security assessments and supervision, propose rectification requirements, and set deadlines. Providers shall retain logs and cooperate by providing technical and data support.
Articles 29–30 — Confidentiality and Public Complaints
Institutions and personnel involved in assessments shall keep personal privacy, personal information, and trade secrets confidential (Art. 29). Any organization or individual may report violations to cyberspace and relevant departments, which shall handle reports promptly (Art. 30).
Chapter V — Legal Liability
Article 31 — Penalties for Core Service and User-Rights Violations
Violations of Articles 7, 8, 9(1), 10, 14, 16, 17, 22, 24, and 26: where laws or administrative regulations apply, penalties shall be imposed accordingly. Where no specific provisions exist, the CAC and relevant departments shall issue warnings, public criticism, and orders for rectification; where correction is refused or circumstances are serious, suspension of information updates shall be ordered, with a fine of RMB 10,000 to RMB 100,000. Criminal violations shall be pursued accordingly.
Article 32 — Penalties for Content Standards and User-Protection Violations
Violations of Articles 6, 9(2), 11, 13, 15, 18, 19, 20, 21, 27, and 28(2): competent departments shall impose penalties in accordance with applicable laws, administrative regulations, and departmental rules.
Article 33 — Fraudulent Filing and Revocation
Where filing is obtained through concealment or false materials, authorities shall revoke the filing, issue a warning and public criticism; where circumstances are serious, suspension of information updates and a fine of RMB 10,000 to RMB 100,000 shall be imposed.
Filing shall be cancelled where a provider terminates services without proper deregistration, or where it has been subjected to serious legal sanctions such as closure, licence revocation, or business licence revocation.
Chapter VI — Supplementary Provisions
Article 34 — Interpretation Authority
These Provisions shall be interpreted by the Cyberspace Administration of China together with the Ministry of Industry and Information Technology, Ministry of Public Security, and State Administration for Market Regulation.
Article 35 — Effective Date
These Provisions shall come into force on March 1, 2022.
↑ Back to top
互联网信息服务算法推荐管理规定
（2021年12月31日公布，自2022年3月1日起施行，第9号令）
来源：中国网信网
联合发布机构（4个）
国家互联网信息办公室 · 工业和信息化部 · 公安部 · 国家市场监督管理总局
第一章 总则
第一条
为了规范互联网信息服务算法推荐活动，弘扬社会主义核心价值观，维护国家安全和社会公共利益，保护公民、法人和其他组织的合法权益，促进互联网信息服务健康有序发展，根据《中华人民共和国网络安全法》、《中华人民共和国数据安全法》、《中华人民共和国个人信息保护法》、《互联网信息服务管理办法》等法律、行政法规，制定本规定。
第二条
在中华人民共和国境内应用算法推荐技术提供互联网信息服务（以下简称算法推荐服务），适用本规定。法律、行政法规另有规定的，依照其规定。
前款所称应用算法推荐技术，是指利用生成合成类、个性化推送类、排序精选类、检索过滤类、调度决策类等算法技术向用户提供信息。
第三条
国家网信部门负责统筹协调全国算法推荐服务治理和相关监督管理工作。国务院电信、公安、市场监管等有关部门依据各自职责负责算法推荐服务监督管理工作。
地方网信部门负责统筹协调本行政区域内的算法推荐服务治理和相关监督管理工作。地方电信、公安、市场监管等有关部门依据各自职责负责本行政区域内的算法推荐服务监督管理工作。
第四条
提供算法推荐服务，应当遵守法律法规，尊重社会公德和伦理，遵守商业道德和职业道德，遵循公正公平、公开透明、科学合理和诚实信用的原则。
第五条
鼓励相关行业组织加强行业自律，建立健全行业标准、行业准则和自律管理制度，督促指导算法推荐服务提供者制定完善服务规范、依法提供服务并接受社会监督。
第二章 信息服务规范
第六条
算法推荐服务提供者应当坚持主流价值导向，优化算法推荐服务机制，积极传播正能量，促进算法应用向上向善。
算法推荐服务提供者不得利用算法推荐服务从事危害国家安全和社会公共利益、扰乱经济秩序和社会秩序、侵犯他人合法权益等法律、行政法规禁止的活动，不得利用算法推荐服务传播法律、行政法规禁止的信息，应当采取措施防范和抵制传播不良信息。
第七条
算法推荐服务提供者应当落实算法安全主体责任，建立健全算法机制机理审核、科技伦理审查、用户注册、信息发布审核、数据安全和个人信息保护、反电信网络诈骗、安全评估监测、安全事件应急处置等管理制度和技术措施，制定并公开算法推荐服务相关规则，配备与算法推荐服务规模相适应的专业人员和技术支撑。
第八条
算法推荐服务提供者应当定期审核、评估、验证算法机制机理、模型、数据和应用结果等，不得设置诱导用户沉迷、过度消费等违反法律法规或者违背伦理道德的算法模型。
第九条
算法推荐服务提供者应当加强信息安全管理，建立健全用于识别违法和不良信息的特征库，完善入库标准、规则和程序。发现未作显著标识的算法生成合成信息的，应当作出显著标识后，方可继续传输。
发现违法信息的，应当立即停止传输，采取消除等处置措施，防止信息扩散，保存有关记录，并向网信部门和有关部门报告。发现不良信息的，应当按照网络信息内容生态治理有关规定予以处置。
第十条
算法推荐服务提供者应当加强用户模型和用户标签管理，完善记入用户模型的兴趣点规则和用户标签管理规则，不得将违法和不良信息关键词记入用户兴趣点或者作为用户标签并据以推送信息。
第十一条
算法推荐服务提供者应当加强算法推荐服务版面页面生态管理，建立完善人工干预和用户自主选择机制，在首页首屏、热搜、精选、榜单类、弹窗等重点环节积极呈现符合主流价值导向的信息。
第十二条
鼓励算法推荐服务提供者综合运用内容去重、打散干预等策略，并优化检索、排序、选择、推送、展示等规则的透明度和可解释性，避免对用户产生不良影响，预防和减少争议纠纷。
第十三条
算法推荐服务提供者提供互联网新闻信息服务的，应当依法取得互联网新闻信息服务许可，规范开展互联网新闻信息采编发布服务、转载服务和传播平台服务，不得生成合成虚假新闻信息，不得传播非国家规定范围内的单位发布的新闻信息。
第十四条
算法推荐服务提供者不得利用算法虚假注册账号、非法交易账号、操纵用户账号或者虚假点赞、评论、转发，不得利用算法屏蔽信息、过度推荐、操纵榜单或者检索结果排序、控制热搜或者精选等干预信息呈现，实施影响网络舆论或者规避监督管理行为。
第十五条
算法推荐服务提供者不得利用算法对其他互联网信息服务提供者进行不合理限制，或者妨碍、破坏其合法提供的互联网信息服务正常运行，实施垄断和不正当竞争行为。
第三章 用户权益保护
第十六条
算法推荐服务提供者应当以显著方式告知用户其提供算法推荐服务的情况，并以适当方式公示算法推荐服务的基本原理、目的意图和主要运行机制等。
第十七条
算法推荐服务提供者应当向用户提供不针对其个人特征的选项，或者向用户提供便捷的关闭算法推荐服务的选项。用户选择关闭算法推荐服务的，算法推荐服务提供者应当立即停止提供相关服务。
算法推荐服务提供者应当向用户提供选择或者删除用于算法推荐服务的针对其个人特征的用户标签的功能。
算法推荐服务提供者应用算法对用户权益造成重大影响的，应当依法予以说明并承担相应责任。
第十八条
算法推荐服务提供者向未成年人提供服务的，应当依法履行未成年人网络保护义务，并通过开发适合未成年人使用的模式、提供适合未成年人特点的服务等方式，便利未成年人获取有益身心健康的信息。
算法推荐服务提供者不得向未成年人推送可能引发未成年人模仿不安全行为和违反社会公德行为、诱导未成年人不良嗜好等可能影响未成年人身心健康的信息，不得利用算法推荐服务诱导未成年人沉迷网络。
第十九条
算法推荐服务提供者向老年人提供服务的，应当保障老年人依法享有的权益，充分考虑老年人出行、就医、消费、办事等需求，按照国家有关规定提供智能化适老服务，依法开展涉电信网络诈骗信息的监测、识别和处置，便利老年人安全使用算法推荐服务。
第二十条
算法推荐服务提供者向劳动者提供工作调度服务的，应当保护劳动者取得劳动报酬、休息休假等合法权益，建立完善平台订单分配、报酬构成及支付、工作时间、奖惩等相关算法。
第二十一条
算法推荐服务提供者向消费者销售商品或者提供服务的，应当保护消费者公平交易的权利，不得根据消费者的偏好、交易习惯等特征，利用算法在交易价格等交易条件上实施不合理的差别待遇等违法行为。
第二十二条
算法推荐服务提供者应当设置便捷有效的用户申诉和公众投诉、举报入口，明确处理流程和反馈时限，及时受理、处理并反馈处理结果。
第四章 监督管理
第二十三条
网信部门会同电信、公安、市场监管等有关部门建立算法分级分类安全管理制度，根据算法推荐服务的舆论属性或者社会动员能力、内容类别、用户规模、算法推荐技术处理的数据重要程度、对用户行为的干预程度等对算法推荐服务提供者实施分级分类管理。
第二十四条
具有舆论属性或者社会动员能力的算法推荐服务提供者应当在提供服务之日起十个工作日内通过互联网信息服务算法备案系统填报服务提供者的名称、服务形式、应用领域、算法类型、算法自评估报告、拟公示内容等信息，履行备案手续。
算法推荐服务提供者的备案信息发生变更的，应当在变更之日起十个工作日内办理变更手续。
算法推荐服务提供者终止服务的，应当在终止服务之日起二十个工作日内办理注销备案手续，并作出妥善安排。
第二十五条
国家和省、自治区、直辖市网信部门收到备案人提交的备案材料后，材料齐全的，应当在三十个工作日内予以备案，发放备案编号并进行公示；材料不齐全的，不予备案，并应当在三十个工作日内通知备案人并说明理由。
第二十六条至第三十条
完成备案的提供者应当在其网站或应用程序的显著位置标明备案编号并提供公示链接（第二十六条）。具有舆论属性或者社会动员能力的提供者应当按照国家有关规定开展安全评估（第二十七条）。网信部门会同有关部门对算法推荐服务开展安全评估和监督检查，提供者应当保留日志并配合提供技术、数据支持（第二十八条）。参与评估的机构和人员应依法保密（第二十九条）。任何组织和个人均可向有关部门投诉举报（第三十条）。
第五章 法律责任
第三十一条至第三十三条
违反第七条、第八条、第九条第一款、第十条、第十四条、第十六条、第十七条、第二十二条、第二十四条、第二十六条的，法律有规定的从其规定；无规定的，给予警告、通报批评、限期改正；拒不改正或情节严重的，责令暂停信息更新并处一万元以上十万元以下罚款（第三十一条）。
违反第六条、第九条第二款、第十一条、第十三条、第十五条、第十八条、第十九条、第二十条、第二十一条、第二十七条、第二十八条第二款的，按照有关法律、行政法规和部门规章的规定予以处理（第三十二条）。
通过隐瞒情况或提供虚假材料取得备案的，撤销备案、给予警告；情节严重的，暂停信息更新并处一万元以上十万元以下罚款（第三十三条）。
第六章 附则
第三十四条至第三十五条
本规定由国家互联网信息办公室会同工业和信息化部、公安部、国家市场监督管理总局负责解释（第三十四条）。
本规定自2022年3月1日起施行（第三十五条）。
↑ 返回顶部
