China

AI Safety Governance Framework 3.0

Third edition of China's technical AI safety governance framework. Updates risk categories and technical and institutional responses, including risks from AI agents. Issued by TC260 under CAC guidance; this is a governance framework, not an enacted national statute.

Ethical Guidelines for Artificial Intelligence Medical Imaging Research

Research ethics guidance from the Medical Ethics Subcommittee of China's National Science and Technology Ethics Committee, published by MOST for reference use. Addresses AI medical-imaging research, clinical validation, consent, human decision-making, data security, synthetic-data disclosure, bias and hallucination risks, model generalization and ongoing ethics review. It is an ethics guideline, not a statute or ministerial rule.

China has formulated nearly 200 key standards for AI sector

President Xi Jinping Attends the Opening Ceremony of the 2026 World AI Conference and High-Level Meeting on Global AI Governance and Delivers Keynote Address_Ministry of Foreign Affairs of the People's Republic of China

关于国家标准《网络安全技术 人工智能应用安全分类分级方法》征求意见稿征求意见的通知

关于发布《网络安全标准实践指南——智能体部署使用安全指引》的通知

China unveils measures to boost 'AI plus consumption'

SSE Listing Review Guideline No. 10: Application of the Fifth STAR Market Listing Standard to AI Large-Model Enterprises

Shanghai Stock Exchange business guidance for AI large-model enterprises applying under the fifth STAR Market listing standard. Requires substantiation and disclosure of technical advantages, deployment and commercialisation milestones, relevant approvals and filings, and compliance considerations including data security, privacy and model safety. This is an exchange listing-review instrument, not national legislation; no blanket legal binding classification is inferred.

Implementation Opinions on Artificial Intelligence Plus Information and Communications Innovation (2026-2028)

Official Chinese AI-governance instrument. Normative policy document (规范性文件); non-binding.

US, China opt out of joint declaration on AI use in military

As of today, the REAIM 2026 Summit outcome document has been endorsed by: Albania, Armenia, Australia, Austria, Belgium, Bulgaria, Canada, Chile, Croatia, Cyprus, Czech Republic, Denmark, Ecuador, Finland, France, Georgia, Germany, Ghana, Greece, Hungary, Iceland, Ireland, Italy, Japan, Kenya, Luxembourg, Malta, Morocco, Netherlands, Norway, Pakistan, Panama, Philippines, Portugal, Republic of Korea, Romania, Senegal, Singapore, Slovenia, Spain, Sweden, Switzerland, Turkiye, Ukraine, United Kingdom.​​

2026 Government Work Report delivered at National People's Congress

Premier Li Qiang delivered the 2026 Government Work Report at the Fourth Session of the 14th National People's Congress. The report frames AI as central to China's next phase of economic development and calls for the creation of a new 'intelligent economy' through large-scale deployment of AI agents (智能体) and intelligent computing clusters (智算集群). It reaffirms the 'AI+' strategy launched in State Council Opinions Guofa [2025] No. 11 and sets out renewed commitments to integrate AI into manufacturing, healthcare, education, scientific research, and public administration. The report also stresses the scaling of national computing infrastructure, accelerated domestic development of AI chips, and continued engagement with China's proposed Global AI Governance framework.

China targets quantum leadership and AI expansion in 15th 5 Year Plan

Cybersecurity standard practice guidelines for deployment and use of openclaw-type intelligent agents (No. TC260-PG-2026NA)

On 15 April 2026, the National Information Security Standardisation Technical Committee closes the consultation on cybersecurity standard practice guidelines for the deployment and use of openclaw-type intelligent agents. The guidelines apply to organisations managing such deployments internally, though commercial agent deployments are excluded. The guide sets out security obligations across the full agent lifecycle, covering installation, configuration, usage, and uninstallation, including r...

Ministry of Education and four other departments issue 'AI+Education' Action Plan

The Ministry of Education, jointly with the National Development and Reform Commission, Ministry of Industry and Information Technology, and two other departments, has issued the 'AI+Education' Action Plan (教科信〔2026〕1号). The plan sets out a nationwide initiative to integrate AI across all educational stages — from primary through higher education — covering curriculum design, AI-assisted teaching tools, teacher training, and student AI literacy. It mandates the construction of dedicated educational AI computing infrastructure, the development of high-quality educational datasets, and the deployment of sector-specific large language models. The plan further establishes ethical and safety requirements for AI deployment in schools, prioritises equitable access across urban and rural regions, and integrates AI competency into teacher qualification standards.

人工智能拟人化互动服务管理暂行办法 (Interim Measures for the Administration of AI Anthropomorphic Interactive Services)

Provisional measures governing AI anthropomorphic interaction services, jointly released by five departments (国家网信办、国家发展改革委、工业和信息化部、公安部、市场监管总局). Promotes healthy development and standardizes applications of AI services that simulate human personality traits. Providers must not generate content endangering national security, inciting subversion, or harming minors. Establishes safety management obligations, protections for minors and elderly, personal information handling requirements, security assessments, and algorithm registration. Encourages innovation in cultural dissemination and elderly care under a human-centered principle.

《人工智能拟人化互动服务管理暂行办法》答记者问 (Official Q&A on AI Anthropomorphic Interaction Services)

Official Q&A from China Cyberspace Administration explaining the AI Anthropomorphic Interaction Services regulation effective July 15, 2026. Covers regulation background, scope, governance principles, prohibited activities, service specifications, provider intervention obligations, and protections for minors and elderly. Providers must prevent excessive user dependency, conduct safety assessments for services with over 1 million registered users, and implement special protections for children under 14 including parental consent requirements.

专家解读｜人工智能拟人化互动服务的制度指引 (Expert Analysis: Institutional Guidance for AI Anthropomorphic Services)

Expert interpretation of the AI Anthropomorphic Interaction Services regulation, providing institutional guidance on compliance.

专家解读｜智能向善 构建人机和谐共生的治理新范式 (Expert Analysis: Building Harmonious Human-Machine Governance)

Expert analysis on the governance framework for AI anthropomorphic interaction services, emphasizing principle of intelligence for good and harmonious human-machine coexistence.

Ministry of Industry and Information Technology security guide on cross-border transfer of automobile data (2026)

On 4 February 2026, the Ministry of Industry and Information Technology and 7 other government departments jointly issued the 2026 edition of the security guide on the cross-border transfer of automobile data. The guide aims to support an efficient and secure mechanism for cross-border automobile data flows and to facilitate automobile data exports. Structured in 4 parts, the document sets out general provisions, rules for identifying important data in typical automotive scenarios, the data e...

National Information Security Standardisation Technical Committee Notice recommending 7 national standards for cybersecurity

On 4 February 2026, the Secretariat of the National Cybersecurity Standardisation Technical Committee (TC260) announced a plan covering seven recommended national cybersecurity standards, following the release of a recommended national standards programme by the Standardisation Administration of China. The plan provides for the revision of six existing standards, including those on information security governance, anonymous entity authentication mechanisms based on weak secrets, technical req...

World Data Organisation for global data governance

On 30 March 2026, the World Data Organisation (WDO) was established in Beijing as an international organisation focused on data development and governance practices. The WDO aims to promote alignment across national data policies, develop industry standards and recommendations, and reduce compliance costs for multinational enterprises. Its activities include addressing cross-border data flows, supporting data use in sectors such as healthcare and education, and providing technical assistance ...

ASEAN and China reinforce commitment to deepening comprehensive strategic partnership

Primary source (asean.org) dated 30 March 2026 in CN. See linked source for full text.

Standardisation of short video content labelling

On 21 March 2026, the Cyberspace Administration of China (CAC) announced plans to guide websites and platforms in standardising the labelling of short video content. The announcement outlines three measures, the standardisation of tags used for such content, including clarification of the categories to be applied, the introduction of content labelling as a mandatory step when posting short videos, and the phased review of existing content, including supplementary labelling. The CAC stated tha...

Cyberspace Administration of China releases guidance for online platforms to standardize short video content labeling

当前，网站平台关于短视频内容标注的标准和尺度不一，一些含有虚构演绎、摆拍营销、AI生成等内容的短视频未进行规范标注，严重误导公众认知，扰乱社会秩序，污染网络生态。

Artificial Intelligence Security Standards Working Group (WG9)

Official source record dated 2026-03-20 for China (People's Republic of China) concerning Artificial Intelligence Security Standards Working Group (WG9). See the linked tc260.org.cn source for the authoritative text, procedural context, and implementation details.

AI被“投毒”，如何避免上当受骗

央视“3·15”晚会揭露的AI（人工智能）“投毒”黑产，撕开了生成式AI商业化进程中的灰色地带。当用户对AI的信任度不断提升，将其视为获取客观信息的重要渠道时，一些GEO（生成式引擎优化）服务商却通过系统性投喂虚假信息，让AI沦为商家的“营销傀儡”。

National Cybersecurity Standardisation Technical Committee source on artificial intelligence governance

Official source record dated 20 March 2026 for China (People's Republic of China) concerning National Cybersecurity Standardisation Technical Committee opens solicitation on membership of Artificial Intelligence Security Standards Working Group (WG9). See the linked tc260.org.cn source for the authoritative text, procedural context, and implementation details.

CAC Domestic Generative AI Services Filing List

On 17 March 2026, the Cyberspace Administration of China (CAC) announced that 48 new generative artificial intelligence (AI) services and 46 AI applications or functions completed their respective filing and registration procedures between January and February 2026. These administrative actions were conducted in accordance with the Interim Measures for the Management of Generative Artificial Intelligence Services to promote the regulated application of AI technologies. The Measures require pr...

Guideline on cybersecurity requirements for consumer-type connected cameras with cybersecurity identification

On 17 March 2026, the National Cybersecurity Standardisation Technical Committee closes the consultation on a guideline for cybersecurity requirements for consumer-type connected cameras with cybersecurity identification. The guideline applies to manufacturers and developers of consumer connected cameras used by individuals and households. It outlines cybersecurity capability requirements across five areas: physical and hardware security, system and software security, network and communicatio...

State Administration for Market Regulation assessment of compliance with advertising obligations

On 14 March 2026, the State Administration for Market Regulation (SAMR) announced measures taken to strengthen advertising supervision and enforcement. The measures apply to sectors including financial services and digital advertising formats, including live-streaming and Artificial Intelligence (AI)-generated content. It was highlighted that in 2025, authorities investigated 44,500 violations and imposed fines and confiscations totalling RMB 252 million, over misleading claims, financial ret...

China’s Supreme Court says it treats AI cases with care without stifling growth or innovation

最高人民法院政务网站,最高人民法院网，最高人民法院，最高人民法院网是人民群众了解和联系最高人民法院的重要窗口，承载着司法公开、法治宣传、服务群众、接受监督等重要使命。是最高人民法院的政务网站，是最高人民法院在互联网上唯一的正式身份。

China issues trial measures for ethical review of AI technology

各省、自治区、直辖市及计划单列市、新疆生产建设兵团工业和信息化、发展改革、教育、科技、农业农村、卫生健康、网信主管部门、科协，中国人民银行上海总部，各省、自治区、直辖市及计划单列市分行，中国科学院院属各单位，各有关单位 ：

Cyberspace Administration of China drafts measures for digital virtual human services

为了促进数字虚拟人信息服务健康发展和规范应用，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，国家互联网信息办公室起草了《数字虚拟人信息服务管理办法（征求意见稿）》，现向社会公开征求意见。公众可通过以下途径和方式提出反馈意见：

数字虚拟人信息服务管理办法（征求意见稿）- Management Measures for Digital Virtual Human Information Services (Draft)

Proposed management measures for digital virtual human services in China. Defines digital virtual humans as virtual entities simulated using AI technologies that mimic human appearance, voice, behavior, and interaction. Requires consent for using personal sensitive information in virtual human modeling. Prohibits infringement on portrait rights, reputation, privacy, and IP. Mandates persistent digital human identification labels, data security obligations, risk monitoring mechanisms, user complaint systems, and algorithm filing for services with public influence. Comment deadline: May 6, 2026.

Cyberspace Administration of China guidelines for data classification and grading of financial information services

On 23 February 2026, the Cyberspace Administration of China closes the consultation on the draft guidelines for data classification and grading of financial information services. The aim of the guidelines is to set out how the four-tier level data classification framework of core, important, sensitive, and general data applies to financial information services, with the aim of allowing financial information service providers to comply with cybersecurity requirements. The guidelines thus provi...

Guidelines on Anti-Monopoly Compliance of Internet Platforms

On 13 February 2026, the State Administration for Market Regulation (SAMR) adopted the guidelines on anti-monopoly compliance of internet platforms. The guidelines aim to assist platform operators in identifying antitrust compliance risks and to improve the clarity of the provisions by incorporating lessons from antitrust regulatory enforcement. The document outlines eight distinct risk scenarios for platform operators, presented with examples. These include algorithmic collusion between plat...

China unveils roadmap to integrate AI into public procurement

CAC publishes batch of typical cases regarding investigations into online platforms over allegations of failure to label AI-generated content

近期，部分网络账号发布AI生成合成信息时，不添加AI标识，存在利用虚假不实内容欺骗误导公众的问题，破坏网络生态，造成恶劣影响。网信部门督促网站平台深入排查整治，依法依约处置账号13421个，清理违法违规信息54.3万余条。现将部分典型案例通报如下：

Implementation Opinions on Accelerating Artificial Intelligence Adoption in Public Procurement and Bidding

Official Chinese AI-governance instrument. Normative policy document (规范性文件); non-binding.

Guide on cybersecurity standards-technical requirements for network data labelling

On 4 February 2026, the National Network Security Standardization Technical Committee Secretariat (TC260) adopted the practice guide to cybersecurity standards-technical requirements for network data labelling to assist network data processors in improving security management. The document establishes technical specifications for network data identification, including terminology, attribute formats, generation rules, and log retention requirements. It aims to facilitate the protection of impo...