# China Generative AI Regulations

China's Interim Measures for the Management of Generative AI Services, effective 15 August 2023, regulate public-facing generative AI, balancing innovation with security and socialist values.

The Interim Measures for the Management of Generative AI Services (Chinese: 生成式人工智能服务管理暂行办法; pinyin: Shēngchéng shì réngōng zhìnéng fúwù guǎnlǐ zànxíng bànfǎ) are a set of regulations introduced by China to govern public-facing [generative artificial intelligence](https://www.wikiprompt.org/wiki/generative-ai) services. Issued by the Cyberspace Administration of China and six other national regulators, the measures took effect on 15 August 2023. They represent one of the first comprehensive national frameworks for generative AI, aiming to balance technological development with national security and social stability.

The measures apply to services provided to the public within the People's Republic of China, exempting internal use by companies and research institutions. They establish requirements for data quality, content labeling, user privacy, and supervision, while encouraging innovation and international cooperation. The framework reflects China's approach of "inclusive prudence" - fostering growth while imposing clear compliance obligations.

## Background and Issuance

The rapid advancement of [large language models](https://www.wikiprompt.org/wiki/large-language-model) and [generative AI](https://www.wikiprompt.org/wiki/generative-ai) systems in 2022 and 2023 prompted regulatory responses worldwide. China's Cyberspace Administration (CAC) led the drafting, collaborating with the National Development and Reform Commission, Ministry of Education, Ministry of Science and Technology, Ministry of Industry and Information Technology, Ministry of Public Security, and National Radio and Television Administration. The multi-agency issuance signaled a coordinated approach across technology, education, and security domains.

The measures were published in July 2023 and took effect on 15 August 2023, following a period of public consultation. They build on earlier regulations for algorithms and deepfakes, extending oversight to generative AI specifically. The timing coincided with the global launch of several commercial [generative AI](https://www.wikiprompt.org/wiki/generative-ai) products, making China one of the first major economies to enact binding rules.

## General Provisions

Article 1 states the measures' purpose: promoting healthy development and regulated usage of generative AI while safeguarding national security, public interests, and citizens' legal rights. Article 2 defines the scope, applying to public-facing services within China, but exempting internal services by organizations that do not serve the domestic public. This exemption allows companies to use AI internally without the same compliance burden.

Article 3 establishes guiding principles: equal emphasis on development and security, promoting innovation alongside rule of law, and adopting "inclusive prudence" with classified supervision. This reflects a pragmatic stance, avoiding overly restrictive rules that could stifle innovation. Article 4 requires services to adhere to laws and social morality, specifically aligning with Core Socialist Values. It prohibits content that promotes overthrowing the socialist system, harms national security, or undermines stability, as well as discrimination, terrorism, extremism, violence, obscenity, and false information. Models must be designed to avoid algorithmic bias, respect intellectual property, and protect users' physical and mental health.

## Technology Development and Governance

Articles 5 and 6 encourage innovation across industries, supporting collaboration on AI research, data, and applications. They promote independent innovation in core technologies, infrastructure, and data resource platforms, while encouraging international cooperation and sharing of computing resources. The measures also call for expanding high-quality public training data and adopting secure chips - a nod to concerns about supply chain security.

Article 7 mandates that providers ensure training data is high-quality, accurate, diverse, and legitimately sourced, without infringing intellectual property rights. If personal information is used, explicit user consent is required. This provision addresses issues like copyright infringement and privacy violations in AI training. Article 8 requires providers to establish clear data labeling rules, including quality assessment, sampling verification, and personnel training, to ensure reliability.

## Service Specifications

Article 9 legally designates providers as producers of online content and processors of personal data, requiring them to sign service agreements with users. This clarifies legal responsibilities. Article 10 addresses minors, mandating effective measures to prevent excessive use and requiring providers to explain appropriate usage. Article 11 obligates providers to protect user privacy, minimize data collection and retention, and allow users to review, copy, modify, or delete their personal information.

Article 12 requires labeling of synthesized pictures, videos, and other generated content, a key transparency measure to prevent deception. Article 13 mandates stable and continuous services for normal usage. Article 14 obligates providers to promptly address illegal content or misuse, including halting generation, retraining models, warning users, and reporting to authorities. Article 15 requires complaint and reporting mechanisms for public grievances, with timely processing and feedback.

## Supervision and Legal Liability

Articles 16 through 21 outline oversight and enforcement. Article 16 assigns regulatory roles to cybersecurity, technology, industry, education, public security, and other government bodies. Article 17 requires security assessments for services with public opinion influence, along with algorithm filing procedures - similar to existing algorithm regulations. Article 18 allows users to complain to authorities about non-compliant services. Article 19 enables official inspections, requiring provider cooperation. Article 20 allows the government to address foreign services that violate the measures, extending extraterritorial reach. Article 21 defines authorities that may punish violators, with crimes investigated according to law.

## Supplementary Provisions

Article 22 defines key terms: generative AI, service providers, and service users. Article 23 requires obtaining relevant licenses where required by law. Article 24 establishes the effective date of 15 August 2023. These provisions ensure clarity and legal certainty for implementation.

## Impact and Reception

The measures have been seen as a balanced approach, encouraging innovation while imposing compliance costs. They have influenced global discussions on AI regulation, with other jurisdictions considering similar frameworks. Some experts note that the requirements for data quality and content labeling may increase operational burdens for providers, but the exemption for internal use provides flexibility. The emphasis on international cooperation suggests China aims to participate in global AI governance while maintaining domestic control.

As of 2025, the measures remain in effect, with ongoing updates and related regulations, such as the Interim Measures for the Management of Anthropomorphic AI Interactive Services, reflecting evolving policy. The framework has been praised for its clarity but criticized for potential overreach, particularly regarding content restrictions. Nonetheless, it represents a significant step in the global effort to regulate [artificial intelligence](https://www.wikiprompt.org/wiki/artificial-intelligence).

---
Source: https://www.wikiprompt.org/wiki/china-generative-ai-regulations
License: CC BY-SA 4.0 (https://creativecommons.org/licenses/by-sa/4.0/)
Last updated: 2026-09-12T16:23:14.908919+00:00