LONDON — China will be invited to join a new global AI research body which Rishi Sunak hopes to establish at next month’s summit at Bletchley Park, a U.K. official involved in discussions has told POLITICO.
According to a draft communiqué, Sunak is seeking backing for a global network of AI safety researchers that will support policy making across the world. Britain wants the network to pull on existing research to produce an annual report that will be written by an expert panel chaired by an academic from outside the U.K.
“It is proposed this network should be an informal and inclusive arrangement for supporting bilateral, plurilateral and multilateral research collaboration … it will not make policy or regulatory recommendations, but we hope its outputs can help inform both international and domestic policy-making in this space,” reads the draft document, dated October 16.
Britain is aiming to get backing from a broad range of countries for the network ahead of its AI safety summit on November 1 and 2, including the United States, European Union and China. The new network will be different from existing research initiatives, such as the Franco-Canadian led Global Partnership on Artificial Intelligence (GPAI), because it will involve Chinese participation.
“We’re keen to engage with China on scientific collaboration. They’re not part of GPAI and we’d like to fill that gap,” said the U.K. official quoted above. They were granted anonymity to discuss multilateral discussions.
“Nothing has been agreed,” said a spokesperson for the U.K. government. “Discussions at the AI Safety Summit will involve exploring areas for international collaboration on frontier AI safety and research. There are different ways to do this, which we’ll continue to discuss with partners and industry.”
Britain has invited the Chinese minister for science and technology Wang Zhigang to the summit, said a second U.K. official who was granted anonymity as they were not authorized to speak publicly.
Frictions loom over talks
The new research network will be distinct from the “AI Safety Institute” that Britain will pitch at the summit as a body that will inspect and assess the national security risks associated with the most advanced forms of the emerging technology.
The Rishi Sunak government will discuss the institute on the second day of the summit, which will only involve like-minded countries — so not China — and a smaller group of companies, such as DeepMind, OpenAI and Anthropic, that produce the most advanced forms of AI.
The new network, however, forms part of a communiqué outlining the risks and opportunities of AI that Britain is seeking broad backing for, including from rivals like the U.S. and China.
Details of the broad-based research network emerge as China launched its own “Global AI Governance Initiative” this week. Chinese officials warned about the risks of AI falling into non-state actor hands, the need to develop human-centric AI and how to assess the technology, indicating areas of convergence with the West on high-level principles.
“This has the potential to show clearer lines on what the West and China can work on together on AI,” Sihao Huang, an AI governance expert at Oxford University. “I think there is room for collaboration.”
But the Chinese proposal also sketches out areas of potential friction with the West on AI.
Pitching itself as sticking up for the little guy, Beijing says that “all countries, regardless of their size, strength, or social system, should have equal rights to develop and use AI,” that AI tech should be made available to the public under “open-source terms.”
In a nod to tightening U.S. curbs of exports of the microchips vital for AI development to China, the document says it opposes “creating barriers and disrupting the global AI supply chain through technological monopolies and unilateral coercive measures.”