Xi calls for international cooperation on AI at Shanghai conference

China's President Xi Jinping said artificial intelligence should not be dominated by one country and called for international cooperation on AI development at the opening of the World Artificial Intelligence Conference in Shanghai on Friday.

Xi said AI development should be a cooperative international effort rather than what he called a solo performance by one country. He also said countries should oppose what he described as overstretching the national security concept in AI or putting one country's security above that of others.

The comments came as Chinese AI models are gaining global users with lower costs and narrowing the gap with leading United States models. The source said AI governance has become a key question because of concerns about military use of AI and possible use by hackers and terrorists.

The United States and the European Union restrict technology exports to China over national security concerns. The source also said tensions between Washington and American AI labs have raised questions about who controls leading technologies.

Shengyun Lu, founder of the Shanghai consultancy Praxis Advisory, told AFP that China was trying to lead in both AI technology development and governance. He said he believed AI should be regulated in a way similar to nuclear power.

Xi told the conference that laws, technological monitoring, early warning systems and emergency response systems should be used to help ensure AI remains under human control. He called for a people-centred approach.

On Thursday, China's foreign minister Wang Yi and representatives from 29 countries, including Russia, Pakistan and Indonesia, agreed to establish an intergovernmental AI cooperation group. State media reported that the World Artificial Intelligence Cooperation Organization will be headquartered in Shanghai and will aim to support what it described as the healthy and orderly development of AI.

Lu said he understood why Western countries were absent from the initiative, saying Europe already had its own AI act and the United States was defining its regulations. New York University professor Arun Sundararajan said recent signs of cooperation between Presidents Xi and Trump were encouraging, but that it was hard to imagine a single global approach to AI governance.

The four-day Shanghai conference brings together more than 1,000 Chinese technology firms, officials, researchers and industry figures. About 3,000 products are on display, including semiconductor systems for AI computing and a smartphone that can operate apps autonomously.

Several technology products are being highlighted at the event. Beijing-based Moonshot AI released a new flagship model, Kimi K3, which it said showed frontier-level performance. Other highlighted systems include MiniMax's M3 model and Huawei's Atlas 950 supernode, an AI architecture for learning and reasoning.

Poe Zhao of the analysis publication Hello China Tech said the main theme would be the move from AI models to systems that can be deployed at scale in everyday life.

State media, citing officials, said daily consumption in China of AI tokens, an industry measure of AI use, has increased a thousandfold over the past two years. The source said more companies abroad, including Siemens, are adopting Chinese open-source AI models because of their performance, lower cost and ability to be customised, compared with closed systems from United States companies such as OpenAI and Anthropic.

The source said OpenAI and Anthropic temporarily withheld releases of their latest models because of government concerns that they could help hackers break into critical online infrastructure.

Mike Luan, who works for an AI research lab, told AFP outside the conference venue that models are controlled by very few groups, and said more cooperative governance could expand access to them.