중국의 시진핑은 AI 위험을 공개적으로 인정한 반면, 트럼프 행정부의 미국은 AI 규제를 경시하고 있으며, AI 칩에 대한 상이한 수출 통제 입장을 드러내고 있다.

After chief executives at top artificial intelligence companies called on the U.S. government to cooperate with China to contain the dangers of the technology, President Donald Trump began his summit with Chinese President Xi Jinping by suggesting that he wasn't interested.

"Super Intelligence (SI) will be a big topic of discussion, but I want to leave it exactly where it is," Trump wrote in a post on Truth Social on Thursday, using his preferred term for AI. The president did not mention the technology in his brief remarks at the White House later that day before his talks with Xi.

But when Xi spoke after Trump, the Chinese president brought up AI and said the technology needed to be stewarded responsibly. "Both China and the United States are leading nations in artificial intelligence," Xi said. "We have both the capability and responsibility to develop and manage AI for good, and ensure that the development of AI is always under human control and serves the well-being of the people."

Trump and his allies in Silicon Valley have aggressively rejected recent calls from some AI executives for a slowdown in AI development by arguing that China would respond by racing ahead of the United States. But public pronouncements from Xi and other Chinese officials suggest the picture is more nuanced. Although the idea of a slowdown has been met with skepticism by commentators in China aligned with the government, Xi and other prominent voices are beginning to express concern about the technology becoming dangerous or disruptive to society.

Xi said in a July speech at China's major annual AI conference that oversight of the technology should be developed to "forestall loss of control." The country already has stricter regulations on AI than the U.S. Earlier this month, an official with China's main AI regulation agency warned in a news conference that the increased abilities of AI models to find and exploit software vulnerabilities were giving rise to the risk of "extreme loss of control."

Both presidents signaled in advance that AI would be on the agenda of Thursday's summit. A summary of the meeting released by China said that Xi had told Trump that the two nations "can continue their dialogue on AI, exchange views on its risks and benefits, and jointly prevent the misuse and abuse of AI." It did not mention any discussion of a formal agreement on the technology.

Trump's opposition to AI regulation and China's resentment at U.S. efforts over the past ten years to hurt its AI industry through export controls make any broad U.S.-China agreement on the technology unlikely, experts say. "Reality is going to impose a sense of urgency on both sides," said Sheehan, who researches global tech issues with a focus on China. "Things could change a lot."

U.S. and Chinese leaders have cited artificial intelligence as central to their rivalry for about a decade. Beijing said in 2017 that it aimed to become the world leader in the technology by 2030. Pentagon planners expect AI to play a bigger part in warfare, by enabling autonomous drones and speeding up tactical decisions, potentially upending U.S. military influence in the Pacific. The rapid expansion of AI capabilities has convinced business and political leaders in both countries that the technology will soon be a cornerstone of economic growth, pushing up productivity and leading to new scientific breakthroughs.

Chinese AI companies are highly adept, launching AI models that nearly match the best from American companies, despite U.S. export controls cutting off China from open access to the high-end computer chips needed to train and run advanced AI. The Chinese government has stood behind its domestic AI champions and urged them to become independent of U.S. technology. U.S. corporate leaders have repeatedly said they must race to make AI powerful to stay ahead of China's evolving prowess. Dario Amodei, the CEO of Anthropic, who recently proposed an industry-wide slowdown over safety fears, said in a recent essay that U.S. AI development should not slow so much that China could catch up.

China has implemented a number of national regulations on AI: it has enacted rules requiring AI-generated content to be clearly labeled, limited the use of algorithms in disseminating news and other content, and banned minors from using chatbot companions. Chinese government officials regularly discuss AI and its risks in public speeches and articles, according to people working on AI policy in China.

The same official who publicly warned about the risk of losing control of AI earlier this month also said that the technology was "reducing demand for traditional labor." In a September 12 article published in a Chinese cybersecurity journal, China's minister of state security Chen Yixin warned that propaganda and deepfakes generated by AI may be used to cause political divisions inside China, threatening the country's stability. The Chinese Communist Party must control the technology to minimize these risks, he wrote. "But technical loss of control is definitely on the radar" in China, said Wang, who writes the Pekingnology newsletter.

An increasing number of U.S. politicians in Congress and in state governments from both major parties are talking about the risks of AI. Many have responded to Americans' concerns about the technology's impact on society and the recent debate about safety by pushing for new regulation. Liccardo said in an interview Thursday that it was clear some members of Trump's Cabinet also support more action on AI risks. Treasury Secretary Scott Bessent spoke to his Chinese counterpart this week about setting up a "notification mechanism" between the two countries to keep track of AI-related security incidents.

"I don't know why they need a red phone unless you thought there was a risk here to discuss," Liccardo said. He thinks it's possible that Trump may change his mind on AI in coming months, as the president comes under pressure from some of his advisers and the public's skepticism of the technology continues to grow.

When the White House does decide to act, China—a world player in AI whose technology is used around the world including in the U.S.—will need to be part of its response, Liccardo said. "We have to have China at the table or AI safety will simply not be possible."