In response to China's ambassador to the UK, academics have questioned the openness of China's AI ecosystem, advocating for shared standards and a unique British approach.
GeoGPT's openness questioned
Ambassador Zheng Zeguang celebrated openly released AI models, but Prof Paul H Cleverley from Robert Gordon University argues that China's claim to openness is not fully substantiated. He cites GeoGPT, a geoscience system from Zhejiang Lab, which, despite being promoted as open, does not meet the model openness framework endorsed in a recent UN report. While it releases model weights based on Alibaba's Qwen, it does not release training data or application source code, and its governance is controlled by Zhejiang Lab.
Comparison with other initiatives
Cleverley points to the European Space Agency's Earth Virtual Expert (EVE) and the US non-profit Ai2's OLMo as examples of fuller openness, as they release training and evaluation tooling or everything needed for reproducibility. He emphasizes that a free hosted service is not the same as an open one, as users' data may be subject to China's legal jurisdiction, creating dependency and sovereignty risks.
Call for shared standards
Cleverley notes that these concerns are not unique to China, as the US private-enterprise model raises similar issues. He suggests that the UK should press for shared openness standards, such as the UN-recommended framework, as the basis for cooperation.
Britain's distinctive path
Dr Claire Jenkins from Wollaton, Nottinghamshire, argues that Britain should not see AI as a geopolitical race but focus on building AI that is trustworthy, transparent, and directed towards the public good. She highlights the NHS and public institutions as strengths, and stresses the importance of governance and human oversight in healthcare AI. Jenkins believes Britain could offer a distinctive path combining innovation with democratic accountability, scientific excellence, and social justice.
