China's Ministry of State Security has issued warnings about the security vulnerabilities linked to AI relay services that enable local developers to access foreign AI models, citing risks of data leaks, privacy breaches, and unauthorised data transfers in a budding grey market.
- Relay services aggregate access to multiple foreign and domestic AI models.
- Some platforms retain unencrypted user data risking leaks or illicit trading.
- Backdoors in relay platforms may enable unauthorized system access.
What happened
China's Ministry of State Security (MSS) publicly warned about the growing security risks associated with artificial intelligence relay services that provide Chinese developers with access to foreign AI models. These relay platforms serve as intermediaries, offering a convenient and cost-effective way to use restricted overseas AI systems, including US models that are officially unavailable in mainland China.
The MSS detailed that some relay service operators operate without proper qualifications and weak security measures, increasing the chance of privacy breaches and data leaks. They highlighted scenarios where user data is stored on servers without adequate encryption, potentially leading to unauthorized data sales or surveillance via hidden backdoors embedded in the platforms.
Why it matters
China’s reliance on AI relay services demonstrates the ongoing demand among local developers to access advanced foreign AI models that currently outperform domestic alternatives in specific areas like coding capabilities. Top US AI systems from OpenAI, Anthropic, and Google continue to lead benchmarks, while Chinese models lag behind, incentivizing use of relay services.
The security warnings reflect broader concerns about data sovereignty, national security, and the integrity of AI supply chains. Weak controls on relay platforms could expose sensitive user information to unauthorized parties, and backdoors might allow surreptitious monitoring of developer activities. This jeopardizes trust and may complicate China’s ambitions to develop and compete with cutting-edge AI technologies domestically.
What to watch next
It will be important to monitor the regulatory response in China towards AI relay platforms, including potential crackdowns or requirements for operator qualification and enhanced security standards. The Ministry of State Security’s advisory to select authorized platforms and safeguard user data indicates that tighter controls may be forthcoming to curb grey market activities.
Internationally, US and allied governments remain alert to attempts by China-linked entities to access advanced AI systems via proxy accounts and networks, reflecting ongoing tensions over technology transfer and cybersecurity. Developments in AI benchmarking and model performance could also influence Chinese developers’ access strategies and the role of relay services in bridging domestic and foreign AI technologies.