China’s approach to regulating artificial intelligence (AI) reflects its ambition to lead in both innovation and governance. With the rapid expansion of AI technologies across industries, the government has introduced comprehensive laws and draft standards to address the ethical, security, and operational challenges posed by AI. This article explores key regulations, including the Provisional Measures for the Administration of Generative Artificial Intelligence Services (2023) and the emerging Cybersecurity Technology – Basic Security Requirements for Generative AI Services (2024 Draft), alongside initiatives aimed at promoting AI application across industries.
- Provisional Measures for Generative AI Services (2023)
The Provisional Measures for the Administration of Generative AI Services, effective August 15, 2023, provide comprehensive guidelines for the development and use of generative AI in China. These measures target AI systems that generate content such as text, images, and video, and regulate their use to ensure alignment with national security, public interest, and ethical standards.
Key provisions include:
- Content Control: Generative AI content must reflect socialist core values and not promote subversion, separatism, or false information. The law also mandates that AI-generated content respect intellectual property, privacy, and avoid discrimination.
- Data Protection: The measures emphasize data compliance, requiring that AI service providers use data from lawful sources and secure personal information in line with China’s Personal Information Protection Law (PIPL). Providers must also ensure the integrity and security of training data.
- Service Accountability: AI providers must maintain transparency about their services, disclose how AI systems are trained, and take responsibility for ensuring that their services do not cause harm. Reporting mechanisms must be in place to handle illegal content.
- Innovation-Focused AI Regulations: Demonstration Scenarios for New-Generation AI (2022)
While strict in its regulatory approach, China also encourages AI innovation through strategic initiatives. The 2022 Notice on Supporting the Development of Demonstration and Application Scenarios for New-Generation AI aims to foster AI’s integration into key sectors of the economy. It outlines ten demonstration scenarios, focusing on areas where AI can drive significant societal and economic impact:
- Smart Agriculture: AI technologies like swarm intelligence and unmanned machinery are transforming China’s agricultural sector by optimizing planting, management, and harvesting processes.
- Industrial Automation: AI-driven systems in smart ports and smart factories improve logistics and manufacturing efficiency, enhancing China’s industrial capabilities.
- Healthcare and Education: AI applications in healthcare aim to provide smart diagnosis and improve access to medical services in rural areas. In education, smart classrooms are being promoted to enhance learning outcomes for students in underdeveloped regions.
- Emerging Legal Standards: Cybersecurity Technology – Basic Security Requirements for Generative AI Services (2024 Draft)
The recently proposed Cybersecurity Technology – Basic Security Requirements for Generative Artificial Intelligence Services (Draft for Comment), issued in May 2024, signals the future of China’s AI regulation. This draft regulation, currently open for public comment, aims to establish a cybersecurity baseline for generative AI services, focusing on the entire lifecycle of AI model development, from data collection to content generation. Key provisions include:
- Data Security: The draft outlines stringent requirements for training data security, mandating that service providers conduct security assessments of data sources and filter out illegal or unhealthy content. The standard emphasizes traceability, ensuring that training data, whether self-collected, commercially obtained, or sourced from open platforms, meets strict legal and ethical guidelines.
- Model and Content Security: Providers must ensure the security of their AI models, including regular security audits of AI development frameworks and protection against cyberattacks such as backdoor and injection attacks. In addition, generated content must align with scientific standards and reflect societal norms, avoiding false or misleading information.
- Transparency and User Protection: The draft also requires transparency in AI service operations. Providers must disclose information about the models used and ensure that users can easily terminate the use of their data for training purposes. Furthermore, services that are not appropriate for minors must implement measures to restrict access.
- Legal Trends and Regulatory Focus
China’s regulatory landscape for AI reflects a clear trend: encouraging innovation while maintaining stringent oversight to address ethical, security, and societal concerns. The Provisional Measures and 2024 Cybersecurity Draft emphasize the need for transparency, data protection, and content control, reflecting China’s commitment to both technological leadership and public safety.
Moreover, the 2022 AI Demonstration Scenarios highlight how the government is actively promoting AI in key industries such as healthcare, education, and agriculture. These initiatives reflect a strategic effort to drive both economic growth and social development through AI innovation.
Conclusion
As China continues to advance its AI capabilities, its regulatory framework is evolving to balance innovation with accountability. The Provisional Measures and the Cybersecurity Technology Draft are part of a broader effort to ensure that AI technologies develop responsibly, with due regard for national security, public welfare, and ethical considerations. As these regulations take shape, they will provide a clearer path for AI providers and innovators navigating this dynamic and increasingly regulated landscape.
