Beijing —
While US President Donald Trump accused China of misusing US election data in a televised speech from Washington, China, on the other side of the world, Xi Jinping was delivering a very different message.
“Beijing is a responsible global leader, keen to shape the future of technology for good,” President Xi told hundreds of technology executives, researchers and industry officials gathered in Shanghai on Friday for the opening of China’s flagship artificial intelligence summit.
“As AI advances at breakneck speed, we must ensure that its development is positive, good, and for the benefit of humanity,” Xi said in his opening remarks at the conference. “We must make that oversight and governance accurate and effective, and constantly refine measures to prevent loss of control.”
Mr. Xi’s remarks came minutes after President Trump made a series of allegations against the Chinese government, including that the Chinese government fraudulently obtained 220 million U.S. voter files in a broader effort to influence U.S. elections. China denies the allegations.
The juxtaposition of these two messages highlights how the rapid rise of AI is only deepening the fault lines and anxieties in the technological race between the United States and China.
Xi’s message is a clear call for China to take the lead in setting global rules around AI, and comes as competition between the US and China for the technology intensifies, as well as strong concerns about national security implications, including AI’s ability to exploit vulnerabilities in software and databases.
In his speech, Xi pushed back against “overextending the concept of national security in the field of AI” and “prioritizing the security of one country over the security of others,” hinting at how Beijing views the US approach to the technology.
Instead, China is trying to push the message that this technology should be a “global public good” and that it is ready to work with countries to co-develop the technology.
On the eve of the conference, China launched the World Artificial Intelligence Cooperation Organization (WAICO), a new grouping of 29 countries, including Russia, Indonesia and Pakistan, friendly to China and its objectives.
“Xi sees AI as an opportunity to have more allies to compete with the United States not only in AI technology but also in international relations. This is AI diplomacy,” said George Chen, head of digital practice at Hong Kong-based consultancy Asia Group.
He added that he feels that China has missed the opportunity to set the rules for the global development of the World Wide Web in recent decades, but the emergence of AI has made China’s position stronger.
“Thirty or 40 years ago, China was a very poor country… But we all know it’s different today. If AI is the new internet, China doesn’t want to miss the opportunity again.”
American companies are widely seen as racing to the technological frontier as a core strategy for competitive advantage. Their models still largely hold the lead, not only in functionality, but also in the hardware used to train and advance the models.
However, the gap is narrowing. And when it comes to winning the AI race, experts say the Chinese government is betting on a different strategy. This is the application and expansion of AI technology in robotics and automation, and its large-scale introduction on a global scale.
Chinese artificial intelligence companies such as DeepSeek and Zhipu have made significant strides to close the performance gap with U.S. companies.
Additionally, more and more users around the world are choosing the model’s open source format and lower operating costs compared to Silicon Valley offerings.
On OpenRouter, a platform that allows users to interact with a variety of models, Chinese companies accounted for 20 of the daily top 50 AI models in May, up from just five in early 2025, according to analysis by Our World In Data. Most of the others are American.
The U.S. government has claimed in recent months that Chinese companies are engaged in a “deliberate industrial-scale campaign to distill U.S. frontier AI,” referring to the process by which smaller models train larger models to improve their own capabilities.
Earlier this month, Chinese regulators warned that they had identified significant security “backdoor” risks in U.S. company Anthropic’s Claude Code tool. Anthropic said the so-called backdoor is an experimental mechanism to track abuse of the platform, and access to the backdoor is not allowed in China.
There are also persistent concerns in Washington that foreign adversaries could use powerful AI models to discover and exploit cybersecurity vulnerabilities in America’s critical infrastructure. The White House launched an effort to address them earlier this week.
Reuters reported earlier this month, citing sources, that the Chinese government is also considering the possibility of restricting foreign access to China’s cutting-edge AI models.
The two countries agreed to start an AI dialogue following the summit meeting between President Trump and President Xi Jinping in Beijing in May.
A stronger foothold for China’s AI models globally could aid China’s ambitions to lead in the adoption and regulation of this technology.
The Shanghai conference showed both its breadth and limitations, with limited involvement from U.S. companies, even though state media reported a record number of participants at this year’s event.
Organizers say attendees at the four-day Shanghai conference will include United Nations Secretary-General António Guterres, nine Nobel laureates, the Turing Prize in Computing, and more than 1,000 global companies.
This is the first time Mr. Xi has attended the flagship event since its launch in 2018, underscoring the importance the Chinese government places on AI and the growing competition with the United States to lead its future.
Western analysts have expressed concern that China’s expanded role in setting global standards for AI could allow it to export its highly restrictive media and internet standards.
And there are also questions about how much global interest the Chinese government will be able to generate for its new international organization, WAICO.
Paul Triolo, a partner at DGA Albright Stonebridge Group, a Washington consultancy, said major Western powers are unlikely to sign on to a Chinese-controlled entity that would likely give it broad powers both to advance AI and to control its governance and safety.
“The main action for the United States will be to establish a credible bilateral dialogue with the Chinese government over the governance of frontier AI models,” he said.
“Both sides must grapple with the complex bureaucratic challenges surrounding this issue and the deep mistrust on both sides.”
