context: AI is set to be a core priority in the PRC’s 15th 5-year plan, with a focus on large-scale application and industrial deployment, positioning the PRC at the forefront of application-led AI development. Recent policy signals also suggest a shift in the innovation model—from technology-driven to demand-driven, where markets define needs and technology follows. The rapid rise of OpenClaw reflects this trend, as adoption expands from developers to industry and local governments, which are incorporating it into AI policy frameworks. While central authorities remain cautious over cybersecurity risks and have yet to issue unified rules, the current approach combines local experimentation with limited central oversight.
Zheng Yongnian 郑永年 Chinese University of Hong Kong (Shenzhen) School of Public Policy dean outlined differences between the PRC and the US in AI development, opportunities for international cooperation and key policy priorities.
Zheng characterised PRC–US AI development as complementary rather than purely competitive
- the US focuses more on AGI and frontier models, following an ‘elite route’
- the PRC has advantages in application scenarios and vertical models, following a 'mass route'
- US models are often closed-source and costly, while Chinese models are more cost-effective and application-oriented
- the PRC therefore has greater scope for cooperation with the Global South and Belt and Road partner countries
On cooperation and competition, Zheng noted
- the PRC’s strength lies in real-world deployment and vertical applications
- the US retains an advantage in data integration and general-purpose models
- the two systems are structurally complementary
On the PRC’s AI development priorities, he emphasised
- breaking 'data silos' and enabling orderly data opening
- both data quantity and quality are critical
- the PRC should explore controlled cross-border data flows, balancing security with openness
- leveraging the PRC’s industrial base and application ecosystem
- AI is a 'bottom-up' technology deeply embedded in society
- the PRC’s full-spectrum manufacturing base provides rich deployment scenarios, especially for vertical models
On risks and governance, Zheng highlighted
- AI as an ecosystem rather than just a technology
- applications such as OpenClaw are embedded in social life, requiring broader governance considerations
- data security and differentiated governance
- sensitive data related to national security should remain restricted
- data related to daily life could be opened in a controlled manner
- social and behavioural risks
- AI-driven content and recommendation systems may shape cognition
- safeguards are needed, particularly for younger users, potentially including legislative measures
On foreign investment, Zheng noted
- foreign firms entering the PRC should align with 15th 5-year plan priorities
- similarly, Chinese capital going abroad should align with host countries’ development strategies
- alignment with national planning frameworks is key to smoother market participation