As Trump and Xi Prepare for Summit, How Has AI Become a New 'Battlefield'?

Disputes between the United States and China over how to address the accelerating risks of artificial intelligence are intensifying, as experts from both sides warn that competition between Washington and Beijing is now hindering progress toward a common ground on safety.

Before the highly anticipated summit next week, where observers will closely watch for any signs of cooperation between the two governments in managing AI risks, Chinese think tanks and tech companies have held informal discussions with their American counterparts, involving President Donald Trump and Chinese leader Xi Jinping.

Challenges in Managing Risks

Participants in these talks reported shared concerns on both sides about the dangers posed by extremely powerful AI systems that could potentially go out of control.

Yet they noted that the intense technological competition between Washington and Beijing is making it significantly harder to reach consensus on "regulatory controls" to manage these risks.

These tensions became evident after the CEO of Anthropic, Dario Amodei, published an article late last weekend urging leading U.S. labs to slow down the pace of AI development.

Amodei called for any slowdown to be accompanied by mechanisms ensuring the United States maintains its technological edge over China—particularly through restricting Beijing’s access to advanced chips and countering Chinese companies training their models by exploiting outputs from competing American models, a process known as "distillation".

In contrast, Xiao Qian, deputy dean at Tsinghua University’s Institute for International Governance of Artificial Intelligence, whose department participated in informal talks with U.S. counterparts, expressed disappointment among Chinese experts regarding Amodei’s statements.

"We feel deep concern and disappointment; Chinese AI safety experts wish to work side by side with the United States to manage advanced risks, but Amodei’s article gave the impression that the space for cooperation has narrowed and become more limited than expected," she said.

She added that linking the issue to national security "increases the difficulty of any potential collaboration and makes the landscape more complex."

Point of Conflict

Meanwhile, the Chinese Communist Party-affiliated newspaper Global Times launched a Monday attack on Amodei’s claims that China’s AI progress poses a "serious threat to the United States and the world".

According to the paper, a Chinese expert stated these claims "risk turning AI safety into a zero-sum geopolitical issue, rather than treating it as a global challenge requiring coordination among major technological powers," noting that Amodei’s motivations were "primarily commercial".

With the Trump-Xi summit approaching on September 24, sources familiar with internal U.S. government discussions indicated that expectations for reaching any substantive agreement on AI have sharply declined.

Despite China’s May announcement of a joint commitment by the two leaders to launch an AI dialogue during the Beijing summit, no formal talks have taken place since.

While Trump claimed he discussed "safety safeguards" with Xi, he later stated over the weekend that such safeguards were unnecessary.

Another point of contention emerged in U.S. accusations against Chinese labs for circumventing development rules through the "distillation" technique.

Sources familiar with the matter said U.S. Commerce Department official Bill Gidley raised these concerns during talks with China’s Cybersecurity Administration on the sidelines of the Asia-Pacific Economic Cooperation (APEC) ministerial meeting.

Undermining Prospects for Agreement

In response, Chinese officials rejected these allegations as baseless, emphasizing that they ignore independent innovations and developments by domestic firms.

The White House has publicly warned Beijing against taking strict measures in response to distillation practices; meanwhile, last week, the National Security Agency and other U.S. intelligence agencies accused prominent Chinese entities—including DeepSeek and Alibaba—of conducting distillation operations, stating this was "likely done with knowledge from the Chinese government".

Chinese experts argue that such accusations undermine prospects for agreement.

Sun Chenghao, researcher at Tsinghua University’s Center for International Security and Strategy, stated that "claims attributing the development of Chinese models to distillation from American models create security concerns and political pressures that hinder cooperation in advanced technologies."

Many experts pointed to differing priorities between the two sides regarding the nature of threats: while U.S. discussions focus on future existential risks and preventing AI from going out of control, Chinese organizers emphasize current risks such as cyber intrusions and preventing the spread of politically sensitive content.

A source familiar with Chinese regulatory bodies noted that "imposing strict regulation now does not serve China’s national interest, as the country is still catching up in this field."

Executives and investors in leading Chinese AI groups confirmed there have been no serious or substantive discussions within their organizations about catastrophic or existential risks posed by the technology.

AI Use in Chinese Factories

Peripheral Discussions

On this front, an investor in two Chinese large language model companies said, "We hardly hear any discussion about AI safety at investor meetings."

"Only marginal conversations have recently emerged among researchers and in friendly gatherings, never rising to the level of debate currently underway in the United States; everyone remains focused on enhancing capabilities and solving urgent problems."

A senior executive at a Chinese AI firm noted that prominent developers like DeepSeek and Moonshot aim to achieve general AI, but generally assume their American peers will reach this milestone first.

"The company that develops general AI first will bear the burden of answering ethical questions, while Chinese firms focus on closing the gap and building profitable business models," the executive added.

Nevertheless, concerns about AI are increasingly appearing on China’s political agenda; the Ministry of State Security issued one of the most comprehensive warnings yet from a high-level Chinese authority on the dangers of the technology, citing threats ranging from cyberattacks and data theft to political instability and military conflicts.

A source close to Chinese organizers said this statement sends "a signal to the United States that there is room for dialogue on AI during the Trump-Xi summit," by demonstrating Chinese officials’ interest in addressing technological risks.