China TMT: Bi-monthly Update - March and April 2025 Issue

If you would like to subscribe for our newsletters and be notified of our events on China Technology, Media, and Telecommunications, please contact James Gong at james.gong@twobirds.com.

Key Highlights

In March and April 2025: relevant regulatory authorities advanced the standardized development of key sectors such as artificial intelligence (“AI”), intelligent connected vehicles, telecommunications services, and platform governance through legislation, enforcement, and industry guidance. Meanwhile, protections for minors in the online environment were further strengthened:

• AI: In terms of legislation, the Cyberspace Administration of China (“CAC”) and the National Cybersecurity Standardisation Technical Committee (“TC260”) accelerated efforts to standardize the labelling of AI-generated synthetic content, issuing a series of administrative measures, national standards, and guidelines to instruct relevant enterprises in proper labelling practices. Meanwhile, TC260 further standardized fundamental security requirements for generative AI services, data annotation security, and training data security through national standards. In terms of law enforcement, the CAC continues to implement algorithm filing and administrative penalties, while the Beijing Intellectual Property Court issued the first nationwide ruling affirming the protectability of AI model structures and parameters. Additionally, local governments across China have rolled out action plans and work guidelines to drive the practical implementation and innovation of AI applications across industries.

• Intelligent Connected Vehicles: The Ministry of Industry and Information Technology (“MIIT”) and local industry and information technology authorities issued regulations to standardize product market access, recall management, software over-the-air (“OTA”) upgrade management, road testing, and innovation application management for intelligent connected vehicles.
• Telecommunications Services: The Ministry of Commerce (“MOFCOM”) and MIIT issued policies to advance the opening of telecommunications services and related digital industries, with plans to progressively reduce and eliminate foreign equity restrictions for certain telecommunications services. Meanwhile, the MIIT has launched a number protection service pilot program to mitigate risks of personal information leakage and telecom fraud.
• Platform Governance: The State Administration for Market Regulation (“SAMR”) issued guidance to clarify the compliance obligations of online transaction platforms and published guiding cases to urge online food delivery platforms to conduct qualification reviews of restaurants on the platform. Authorities in Beijing and Heilongjiang intensified enforcement against unreasonable online sales practices and urged platforms to strengthen qualification reviews for restaurants on the platform.
• Protection of Minors: TC260 issued guidelines detailing the technical requirements for juvenile modes on the mobile internet platforms. The CAC also intensified efforts to rectify the improper exploitation of minors’ images and curb the spread of illegal and harmful information to safeguard minors’ rights and interests.

Follow the links below to view the official policy documents or public announcements.

Legislative Developments

1. CAC and other departments jointly released measures to standardise labelling of AI-generated synthetic content (14 March)

The CAC and other departments jointly issued the Administrative Measure for Labelling AI-Generated Synthetic Content to regulate the labelling of AI-generated synthetic content. The measures mandate that generative AI service providers must add explicit labels to text, images, audio, video, and other synthetic content, while embedding implicit labels in file metadata. Users disseminating AI-generated synthetic content through online content dissemination services are required to proactively declare such content and utilize labelling tools provided by service providers. Additionally, application distribution platforms must require application service providers to disclose whether they offer AI-generated synthetic services and verify their compliance with labelling requirements.

2. CAC issued mandatory national standards for AI-generated synthetic content labelling methods (15 March)

The CAC released the mandatory national standard Cybersecurity Techniques – labelling Method for Content Generated by Artificial Intelligence, serving as a supporting document for the Administrative Measures for labelling AI-Generated Synthetic Content, while specifying technical methods for labelling such content. The standard defines explicit and implicit labelling methodologies for AI-generated synthetic content across text, images, audio, video, and other multimodal formats. It also provides reference materials in the appendix, covering labelling approaches, typical application scenarios, labelling examples, and label formats.

3. CAC and CMA issued administrative measures to regulate AI meteorological application services (29 April)

The CAC and China Meteorological Administration (“CMA”) jointly issued the Measures for Artificial Intelligence Meteorological Application Services to regulate AI-driven meteorological services. The measures clarify the rights and obligations of AI meteorological service providers, requiring them to fulfil compliance obligations in accordance with laws, including algorithm filing, security assessments, AI-generated synthetic content labelling, and to establish and improve management policies and technical measures for algorithm security review, cybersecurity, data security, and information release review. Providers are also mandated to implement user-facing complaint and reporting channels.

4. CAC released regulations to govern direct satellite connectivity services for terminal devices (30 April)

The CAC and other departments jointly issued the Regulations on Direct Satellite Connectivity Services for Terminal Devices to govern direct satellite connectivity services for terminal devices. Key provisions stipulate that providers offering direct satellite connectivity services within China must obtain licenses and approvals if terminal devices connect to domestic public telecom networks or use domestic radio frequencies. Providers are also required to comply with cybersecurity, data security, and personal information protection obligations under relevant laws and mandatory national standards.

5. TC260 approved multiple national standards to regulate basic security requirements for generative AI services, data annotation security, and training data security (30 April)

The TC260 approved multiple national standards, including Cybersecurity Technology - Basic Security Requirements for Generative Artificial Intelligence Service, Cybersecurity Technology - Generative Artificial Intelligence Data Annotation Security Specification, and Cybersecurity Technology - Security Specification for Generative Artificial Intelligence Pre-training and Fine-tuning Data. These three standards each focus on distinct aspects of AI security: the Cybersecurity Technology - Basic Security Requirements for Generative Artificial Intelligence Service stipulates fundamental security requirements for generative AI services, including training data security, model security, and security measures, and provides reference points for security assessments; the Cybersecurity Technology - Generative Artificial Intelligence Data Annotation Security Specification specifies security requirements for data annotating platforms/tools, annotating rules, annotating personnel, and annotating verification in generative AI training, and elaborates on security evaluation methods for data annotation; the Cybersecurity Technology - Security Specification for Generative Artificial Intelligence Pre-training and Fine-tuning Data regulates security requirements for processing activities of generative AI pre-training and fine-tuning data.

6. TC260 issued guidelines on coding structures and assignment rules to standardize AI-generated synthetic content labelling (14 March)

The TC260 issued the Practice Guidelines for Cybersecurity Standards – Coding Rules for AI-Generated Synthetic Content Labelling Service Providers to define implicit labelling requirements enshrined in Administrative Measures for Labelling AI-Generated Synthetic Content and Cybersecurity Techniques – labelling Method for Content Generated by Artificial Intelligence. The guidelines provide detailed regulations on the coding structure and coding rules for implicit label in file metadata of AI-generated synthetic content.

7. TC260 issued guidelines detailing technical requirements for juvenile mode on mobile internet platforms (3 April)

The TC260 released the Practice Guidelines for Cybersecurity Standards – Technical Requirements for Juvenile Mode on Mobile Internet, which specifies the technical requirements for mobile smart devices, applications, and mobile application distribution platforms when developing and implementing juvenile modes. The requirements cover functions such as one-click activation and deactivation, parental controls, usage time and duration management, download and installation controls, online addiction prevention, as well as management of external link access and download services. Additionally, the guidelines mandate coordinated efforts among all relevant parties in terms of functional implementation, mode safeguards, and personal information protection to ensure the effective enforcement of the juvenile modes.

8. MIIT and SAMR jointly issued notice to regulate market access, recalls, and OTA upgrade management for intelligent connected vehicles (28 February)

The MIIT and the SAMR jointly issued the Notice on Further Strengthening Market Access, Recalls, and Over-the-Air Upgrade Management for Intelligent Connected Vehicles. The notice aims to standardize market access, recalls, and OTA upgrades for intelligent connected vehicles equipped with combined driver assistance systems and OTA capabilities. It requires enterprises to strengthen capability building and product safety design in accordance with the Technical and Management Guidelines for Market Access, Recalls, and OTA Upgrades of Intelligent Connected Vehicles provided in the annex, and to conduct thorough testing and validation. The notice stipulates that enterprises must report accidents such as collisions that occur during the use of the combined driving assistance system to the MIIT and the SAMR within 48 hours (If the accident results in casualties or has a significant social impact, the report should be made within 24 hours of the accident), using the provided reporting templates in the annex. The notice also stipulates that OTA upgrade activities must be filed with the MIIT and the SAMR.

9. SAMR released guidance to clarify compliance obligations for online transaction platforms (14 March)

The SAMR issued the Guidance on Promoting Compliance Management Responsibilities for Online Transaction Platform Enterprises, aiming to guide platforms in enhancing compliance and fostering healthy development of the platform economy. Key provisions include: (i) strengthening compliance management organization: the opinions suggest that platform enterprises designate chief compliance officers and compliance managers, with network transaction platform principal leader assuming overall compliance responsibility; (ii) clarifying compliance management responsibilities: the opinions further detail the primary responsibilities of  chief compliance officers and compliance managers. (iii) improving compliance mechanisms: the opinions suggest that platform enterprises establish a dynamic management mechanism for compliance risk control, including mechanisms for compliance risk investigation, rectification, management scheduling, and document management.

10. Hangzhou MEIT proposed regulations to refine road testing and innovation applications for intelligent connected vehicles (16 April)

The Hangzhou Municipal Bureau of Economy and Information Technology (“MEIT”) opened public consultation on the Implementation Measures for Innovation Application Management of Intelligent Connected Vehicles in Hangzhou. The measures aim to regulate the entire process of testing and application of intelligent connected vehicles, including the entry conditions for testing and application entities, technical and capability requirements for test vehicles, and procedures for batch confirmation and license plate application. Additionally, the measures require that in the event of a traffic accident involving intelligent connected vehicles, an accident analysis report must be submitted to the municipal authorities within 5 working days.

Enforcement Developments

11. CAC launched “Qinglang Cracking Down on the Misuse of AI Technology” action to clean up and regulate non-compliant AI applications (30 April)

CAC organized a nationwide “Qinglang Cracking Down on the Misuse of AI Technology” action to regulate AI services and applications. The action is divided into two phases: In the first phase, efforts focus on strengthening governance at the source of AI technology, including cleaning up and regulating non-compliant AI applications, enhancing the management of AI-generated synthetic technologies and content labelling, and enhancing platforms’ capabilities to detect and verify authenticity. The second phase targets prominent issues such as using AI to fabricate rumours, spread false information, produce pornographic content, impersonate individuals, or engage in online astroturfing. It will concentrate on clearing related illegal and harmful information and dealing with non-compliant accounts, MCN institutions, and websites and platforms.

12. CAC published tenth batch of deep synthesis service algorithm filing information, covering approximately 400 companies (12 March)

The CAC published a batch of deep synthesis service algorithm filing information, revealing the names of companies, algorithm titles, application products, and their main uses. As of 31 March 2025, a total of 395 deep synthesis service algorithms have completed filing procedures in accordance with the Provisions on the Administration of Deep Synthesis of Internet Information Services.

13. CAC published filing information for generative AI service, covering 346 generative AI services (8 April)

The CAC issued a notice on generative AI service filings from January to March 2025. By 31 March 2025, 346 generative AI services had completed filings with the national CAC under the Interim Measures for the Administration of Generative Artificial Intelligence Services. Additionally, 159 generative AI applications or features that directly invoke filed models via API interfaces or other means were registered with local CAC.

14. SAMR released guiding cases for law enforcement involving third-party online food delivery platforms penalized for failing to review qualifications of restaurants on the platform (April 7)

The SAMR published the first batch of guiding cases for food safety administrative enforcement. Among the seven published cases, three involve online food safety issues, with two cases concerning restaurants that leased licenses and used fake addresses and photos to register on third-party online food delivery platforms. In these two cases, the SAMR ruled that the platforms violated the Measures for the Supervision and Administration of Food Safety in Online Catering Service by failing to verify registered restaurants’ food business license information. The two involved platforms were respectively subject to penalties of confiscation of illegal gains amounting to 106,000 CNY and 459,000 CNY, as well as fines of 200,000 CNY.

15. CAC intensified crackdown on exploitation of minors’ images, penalizing non-compliant accounts in multiple batches (18 April)

The CAC continued to strengthen online protection for minors by removing illegal and harmful content exploiting minors’ images and imposing strict penalties on non-compliant accounts. The identified issues include: (i) inappropriate child parody content for attention-seeking purposes; (ii) staged or fabricated content to provoke controversy; (iii) inappropriate speech distorting public values; (iv) illicit traffic redirection attempts to evade regulatory crackdowns. The CAC stated that they will guide platforms to take strict disposal measures against relevant non-compliant content and accounts in accordance with the principle of prioritizing the best interests of minors, and cooperate with public security agencies to strengthen the fight against illegal crimes.

16. Beijing SAMR issued measures to curb unreasonable business practice in online sales, including algorithmic big data-driven price discrimination and forced bundling (21 April)

The Beijing SAMR issued the Measures to Regulate Unreasonable Business Practices in Online Sales and Live Commerce, targeting three key areas: (i) false advertising: strengthening governance of deceptive promotions and product-misrepresentation violations; (ii) price violations: urging platforms implement clear price labelling and public disclosure policies and banning deceptive pricing tactics; (iii) big data-driven price discrimination and forced bundling: cracking down on algorithmic price discrimination and prohibiting platforms from using technical means such as pre-selected checkboxes or bundled sales to covertly force transactions.

17. Beijing Intellectual Property Court issued landmark ruling affirming protectability of AI model structures and parameters (17 April)

The Beijing Intellectual Property Court concluded the China’s first case affirming the protectability of AI model structures and parameters. In that case, the court decided that defendant’s unauthorized use of the structure and parameters of the plaintiff’s comic transformation filter model violated recognized business ethics in the field of AI models, infringed the plaintiff’s legitimate rights and interests, disrupted market competition order, and harmed long-term consumer interests, thereby constituting unfair competition under Article 2 of the Anti-Unfair Competition Law. The defendant was ordered to pay total compensation of RMB 1.6 million for damages and reasonable expenses.

18. Heilongjiang SAMR launched enforcement action targeting online catering services, urges platforms to implement food waste reduction measures (April 14)

The Heilongjiang SAMR launched the enforcement action targeting online catering services across the province, with a focus on regulating unlicensed restaurants on online food delivery platforms and establishments that have accumulated numerous consumer complaints and negative reviews. Additionally, the action mandates that third-party online food delivery platforms implement anti-food waste measures, provide meal portion alerts to consumers, and supply public utensils and takeaway packaging services upon request.

Industry Developments

19. MOFCOM released work plan to further open up telecommunications services and related digital industries to foreign investment (11 April)

The MOFCOM issued the Work Plan on Accelerating the Comprehensive Pilot Program for Expanding the Opening-Up of the Service Sector, focusing on speeding up and intensifying the pilot opening up of the service sector. The plan explicitly supports opening up telecommunications services and digital industries, including the removal of foreign equity restrictions for internet access services (limited to providing internet access services for users) and information services (limited to app stores, excluding services prohibited for foreign investment sectors). It also opens the domestic internet virtual private network services to foreign investment, with the foreign equity ratio not exceeding 50%. The plan also outlines measures to deepen the opening up of the software and IT service and promote AI industrialization.

20. MIIT proposed guidelines to build comprehensive cloud computing standardization system (22 April)

The MIIT opened public consultation on the Guidelines for Comprehensive Cloud Computing Standardization System Construction (2025 Edition), aiming to strengthen top-level design for cloud computing standardization. The guidelines outline the structure and framework of the system, proposing standards across six areas: foundational standards, technology and product standards, service standards, application standards, management standards, and security standards. By 2027, over 30 new national and industry standards for cloud computing are planned to address needs in foundational technologies, products, services, applications, management, and security.

21. MIIT planned to launch pilot program for number protection services to combat telecom fraud (3 April)

The MIIT opened public consultation on the Notice on Conducting a Pilot Program for Number Protection Services to regulate intermediary number services, strengthen code resource management, and combat telecom fraud and unsolicited commercial messages. The notice designates the 700 number range as dedicated resources for number protection services, with an 11-digit management length and 15-digit usage length, distinct from regular mobile numbers. The pilot will proceed in three phases: a six-month preparation phase, a six-month transition phase, and an 18-month formal pilot phase.

22. Beijing MEIT released action plan to boost AI application capabilities in IT firms (8 April)

The Beijing MEIT issued the 2025 Action Plan to Support IT Software Enterprises in Strengthening AI Application Capabilities, aiming to deepen large-scale AI model adoption and enhance AI’s role in the real economy. The plan outlines eight core tasks: (i) supporting Model-as-a-Service (MaaS) enterprises to cluster in Beijing; (ii) promoting the development of industry-specific model capabilities among information and software enterprises; (iii) promoting general-purpose intelligent agents; (iv) implementing intelligent technological upgrades for IT firms; (v) improving data governance capabilities;(vi) accelerating open-source ecosystem development; (vii) enhancing AI services for small and medium-sized enterprises; (viii) strengthening AI skills training.

23. Shandong issued 2025 work priorities for digital province development, accelerating AI applications innovation (26 April)

Shandong released the 2025 Work Priorities for Digital Province Construction, aiming to establish a digitally advanced province and create an intelligent Shandong. In the field of AI, the document emphasizes strengthening foundational support for AI development and formulating implementation plans to accelerate high-quality AI applications in key sectors. Meanwhile, the document proposes strengthening innovative AI applications and deeply advancing “AI+” scenario applications, focusing on enhancing AI innovation in scenarios such as industrial manufacturing, modern marine affairs, and medical health.

24. Jiangsu launched three-year action plan to drive “AI+” initiatives for new industrialization (19 April)

Jiangsu issued the Three-Year Action Plan for High-Quality Digital Economy Development (2025–2027), focusing on building a data-driven digital economy and a province with strong integration of digital and real economies. The plan includes an “AI+” initiative to empower new industrialization, fostering integrated models like “AI+ low-altitude economy” and “AI+ digital finance”. It also proposes cultivating and introducing basic large model enterprises, supporting leading enterprises to jointly develop and build industry large models, facilitating the establishment of data annotation bases in qualified regions, and developing a full-chain data annotation industry. For key sectors like manufacturing, education, and healthcare, the plan aims to build over 30 high-quality industry datasets and 60+ demonstration application scenarios to facilitate the implementation of large models.

25. Hefei released China’s first data annotation industry development plan to bolster AI development (7 March)

Hefei released the China’s first data annotation industry development plan, the Hefei Data Annotation Industry Development Plan (2025–2027), aiming to build an efficient, intelligent, and secure data industry ecosystem and promote the city’s digital transformation. The plan defines six key tasks: (i) advancing key technologies in data pre-processing, intelligent annotation tools, and data security; (ii) building public service platforms and multimodal crowdsourced intelligent annotation platforms; (iii) cultivating leading enterprises and specialized small and medium-sized enterprises, and improving third-party service systems; (iv) promoting social governance and industrial digital transformation to unleash the value of data elements; (v) improving industry standards, strengthening talent cultivation, and enhancing data regulation; (vi) innovating ecological operations through mechanisms such as government-industry-academia-research integration.

This article was written with the assistance of Derek Xia and Shaun Liu.