The global AI landscape is currently defined by a tension between proprietary development and open-source collaboration. Recently, Chinese Ambassador Zheng Zeguang highlighted the contributions of Chinese labs—specifically those behind models like Qwen, DeepSeek, and Kimi—as evidence of China’s commitment to an open AI ecosystem. These models are noted for their ability to function on modest hardware, which proponents argue provides significant benefits to the developing world.

However, the assertion that openness is a fundamental characteristic of China’s AI strategy is being challenged. Critics suggest that while some Chinese labs have made strides in releasing models, the actual governance and accessibility of these systems remain opaque. The Guardian points to specific examples, such as the GeoGPT geoscience system from Zhejiang Lab, to illustrate that what is presented as a 'jointly governed' open model may not meet the standard definitions of openness expected by the global research community.

Ultimately, the debate reflects a broader geopolitical struggle. While China positions itself as a champion of accessible AI technology, international observers argue that the reality is more complex. The disagreement lies in whether these Chinese initiatives are genuinely open-source or if they are strategic tools designed to project influence while maintaining underlying control. Furthermore, some analysts suggest that this lack of true transparency is not unique to China, but rather a systemic issue across the global AI industry, where major powers frequently conflate marketing claims with actual open-source practices.