| Source | Latest linked headline | Age |
|---|---|---|
| China | The Guardian | CHINA’S AI ECOSYSTEM IS NOT AS OPEN AS IT CLAIMS. NOR IS ANY OTHER COUNTRY’S | LETTERS | 3 hrs ago |
China's ambassador to the UK, Zheng Zeguang, recently argued that openness is a defining feature of China's AI development, citing the release of AI models such as Qwen, DeepSeek, and Kimi. However, according to Prof Paul H Cleverley, this claim deserves scrutiny, as some of these models do not meet the standards of openness. For instance, GeoGPT, a geoscience system from Zhejiang Lab, is promoted as open but does not release training data or application source code.
The GeoGPT model releases model weights built mainly on Alibaba Qwen, whose licenses are not Open Systems Interconnection-compliant, and its governance committee answers to Zhejiang Lab itself. This has led to questions about the true openness of China's AI ecosystem. According to China | The Guardian, this lack of openness is not unique to China, as other countries also have limitations in their AI ecosystems.
The debate around AI openness is significant, as it affects the development and accessibility of AI technology globally, particularly in the developing world. The issue is also relevant to the broader conversation around cybersecurity and technology governance, as seen in related stories about museums facing growing cyber threats and OpenAI putting the brakes on a new model due to its supposed power.