Yi Zeng, a member of the UN Advisory Body on AI, has announced the formation of the Beijing Institute of AI Safety and Governance.
In a tweet on Wednesday, Zeng said “safety and Governance are core capacities for developing healthy and sustainable AI.” He also thanked various Beijing ministries for the support they have given the Institute, of which Zeng will be the founding director.
In October of 2023, The United Kingdom became the first country to launch a governmental institute to tackle the risks of AI, with then Prime Minister Rishi Sunak saying it would “advance the world’s knowledge of AI safety.” Similar institutes have since been formed in the United States, Japan, and Singapore, whilst others agreed to join an international network of AI safety institutes to encourage cooperation between nations. It is not yet known if the Beijing Institute will have a similar structure to the other AI safety institutes, but more details are expected to follow.
The importance of international collaboration on this issue was highlighted by Zeng:
On AI Safety and Governance, industry and academic leaders have to hold hands together for collaborations no matter what, both for the near term and for long-term risks. And we also need to help the governments to understand the necessities to do so.
There has been lengthy discussion about the prospect of US-China cooperation on artificial intelligence over recent years, as AI capabilities (and concerns about the potentially catastrophic outcomes they could bring) have skyrocketed.
It is clear that, like the rest of the world, China takes the existential risk from advanced AI seriously. The only Chinese computer scientist to receive the Turing award, Andrew Yao, and the former president of Baidu, Ya-Qin Zhang, both joined up with Yoshua Bengio (also a Turing Award recipient) and Stuart Russell (author of the most popular textbook on AI) to propose a strategy to mitigate the ‘unacceptable risks’ of advanced AI. Yao said that the world must “work together as a global community to ensure the safe progress of AI”, and Russell argued that AI companies simply should not be able to build their systems if they cannot ensure they will be safe.
Chinese president Xi Jinping has praised Yi Zeng’s work, and has supposedly personally edited a document given to party leaders stating that AI will ultimately determine “the fate of all mankind,” and as such, China should take a proactive regulatory approach to the dangers rather than allowing “uninhibited growth that comes at the cost of sacrificing safety.”