Photo: VCG

China together with some 28 countries and regions, including the US and Europe, were present at the UK AI Safety Summit which was held on November 1 and 2 signing the Bletchley Declaration and the event was splashed all over the headlines.

As a close observer of the AI industry, Zhang Linghan, from China University of Political Science and Law and also one of the only two Chinese experts being enrolled into the High-Level Advisory Body on Artificial Intelligence, sees the Declaration envisions as a global governance mechanism for AI.

The Declaration emphasizes the international nature of the risks posed by AI and proposes the establishment of an international AI safety cooperation mechanism, an international AI safety scientific research network, an international AI safety supervision mechanism, and an international AI safety education mechanism, Zhang said. "It is an important step taken by countries around the world in global AI governance," she added. 

How Zhang views AI governance boils down to cooperation, because whether a cooperative or confrontational attitude is adopted, the risks of AI could spread across national borders. 

Human understanding of AI technology is still lacking, and there are significant differences in the development status of AI and governance experiences regarding AI risks. This has hindered a global consensus being reached.

China, US and Europe, leaders in AI, are facing different industrial situations and governance. Conducting cooperative governance within the scope of consensus is the best choice, Zhang said.

The US recently further tightened the export restrictions on AI-related chips and semiconductor manufacturing equipment to China, and added a number of Chinese entities to the "entity list" of export control.

The US' wishes of slowing down or hindering China's AI technology research and development might work in the short time since the US has dominated the design, manufacturing and sales of advanced graphics computing chips. But ultimately this will only lead to China accelerating its own independent research and development, Zhang noted.

The White House recently issued an executive order on AI which reveals many similarities between the regulatory measures taken by the US and China, such as watermarking for AIGC and requiring companies to conduct prior assessments. These are areas where China and the US can cooperate in the future, Zhang noted. 

China's wisdom 

Previously, whether China would be invited to the summit also grabbed intensive attention. Dust settled when Wu Zhaohui, China's vice minister of science and technology, led the Chinese delegation and delivered a speech at the opening plenary session of the summit.

Regarding what China can offer in AI global governance, Zhang said China is one of the largest AI markets in the world and also has one of the most innovative and dynamic AI research centers. China has made significant development in the field of AI research and actively participated in international cooperation and governance of AI. It is believed that this is also an important reason why the UK invited China to participate in summit.

China's practice of developing and promoting Chinese models has actually provided a lesson for global AI governance - the risks posed by models vary when targeting users of different languages, Zhang emphasized. 

In the field of conversational AI, Chinese content holds an irreplaceable position. On the other hand, Chinese expression also has its unique logic, she said. 

The Chinese expert explained that the risks of domestic AI content may be completely different from foreign AI. Developing Chinese conversational AI requires more attention to screening confusing, ambiguous and sarcastic content. Filtering out "slang" that may involve discrimination and insults is particularly important. 

Natural language models such as ChatGPT and Claude operate in English, and when users engage in multilingual conversations, the models automatically perform cross-language translation, which poses the risk of inaccurate semantic conversion, she added. 

"The greater the difference between the national language and English in terms of language and writing, the more attention needs to be paid to this issue."

The Time magazine citing one expert said "concerns about risks posed by advanced AI are less commonly expressed in China than they are in the West." 

Zhang gave her insight on the issue to the Global Times. She believes that China still plays the role of a follower in the AI industry. From this perspective, it is easier to understand the Western concern about AI risks. 

But that doesn't mean China doesn't attach importance to the security of AI. China's stance has always advocated active participation in AI governance and the promotion of AI technology development, while attaching importance to the protection of national interests, social public interests, and the individual rights of citizens, Zhang said.

In addition, the legal system in China has always focused on the risks brought by AI, and security has always been key word in the country's legal system related to artificial intelligence, the Chinese expert stated. 

Laws and regulations, such as the Data Security Law, the internet information service algorithm recommendation management regulations, the internet information service deep synthesis management regulations, and the interim measures for the administration of generative AI Services, "start from the three elements of data, algorithms, and computing power to construct an AI governance system," Zhang explained. 

Sharing with the world 

"China supports the development of technology for the benefit of people worldwide and is willing to contribute Chinese wisdom to global AI governance." Being the only two Chinese experts from the UN AI body, Zhang told the Global Times she felt incredibly honored to represent China and share its AI governance expertise with the global community.

The development, application, and governance of AI technology are crucial to the fate of all humanity. China conveys its intention of cooperative governance and shared governance to the world through Global AI Governance Initiative, emphasizing that the development of AI technology should adhere to the concept of putting people first and promoting benevolent intelligence, Zhang said.

The concepts and principles proposed in the Initiative, such as the human-centric approach, fairness and non-discrimination, and the principle of sovereign equality, are also mirrored in the "Declaration," reflecting the common needs of the international community, she continued. 
 
China supports exploring global AI governance within the framework of the UN and also looks forward to promoting the formation of governance with Chinese wisdom through multilateral cooperation, Zhang noted. 

Zhang introduced that the UN AI advisory body consist of a total of 39 members, with 19 males and 20 females, representing 33 countries. "As the potential risks and benefits of AI become increasingly evident, the absence of unified governance rules and a coordinated international organization for cross-border AI governance is apparent. With the widely recognized UN global governance core position, now is the opportune moment to commence international governance endeavors for AI."

In the interview with the Global Times, Zhang mentioned that one of the roles of the UN AI advisory body is to offer suggestions to the potential future global AI governance institution. 

Previously, the Financial Times reported that some Chinese AI scientists, along with renowned Western scholars, called for stricter control over AI technology, including the establishment of an international regulatory body, the mandatory registration and auditing of advanced AI systems.

The Financial Times wrote that "the measures are an early indication of China's likely stance on global AI regulation."

Zhang said "It is still too early to talk about what kind of international AI regulatory body can be built as countries are still in negotiations." But what is for sure is that AI governance mechanisms built under the framework of the UN are the most likely to build consensus, the Chinese UN advisory expert said. 

In the past, China has always participated in international AI governance cooperation with a cooperative attitude and will also play an important role in initiatives and proposals put forward by the UN AI body in the future, Zhang noted. Results of a potential future global AI governance institution will be announced at the Summit of the Future in 2024, Zhang said.