The United States and China are set to hold high-level talks on artificial intelligence (AI) in September, as both nations race to establish regulations and address growing tensions in the AI sector. The meeting, expected before Chinese President Xi Jinping’s planned visit to the U.S. on September 24, marks the first U.S.-China dialogue on AI under the Trump administration. Treasury Secretary Scott Bessent will lead the U.S. delegation, though the exact dates and agenda remain unspecified.
The talks come amid escalating concerns over intellectual property (IP) theft and the competitive development of advanced AI models. Bessent warned that the U.S. could impose sanctions on Chinese AI companies if evidence of IP theft is confirmed. He cited concerns that Chinese models, such as Moonshot AI’s Kimi K3, may have been developed using unauthorized access to U.S. AI research, a practice known as 'distillation.'
Kimi K3, released last week, has sparked fears in Silicon Valley and Washington about America’s ability to maintain its lead in AI. The model has been ranked highly in coding benchmarks, rivaling U.S. leaders like Anthropic and OpenAI. Unlike its American counterparts, Kimi K3 is an open-weight model, meaning its training methods are publicly disclosed.
The Trump administration is also in discussions with U.S. AI companies about voluntary safety standards for new AI models. Meanwhile, Congress is considering legislation to impose mandatory safety measures. The administration has previously pressured Anthropic to disable access to its Claude Fable 5 and Mythos 5 models worldwide following a hacking incident linked to China.
Background and Context
The U.S. and China have been locked in a technological arms race, with both nations investing heavily in AI development. The Trump administration has emphasized the need for national security safeguards, including restrictions on American firms using Chinese AI models. Meanwhile, China has accelerated its AI advancements, with Kimi K3 being the latest example of its progress.
Opposing Perspectives
While the U.S. government expresses concerns over IP theft, Chinese officials have not publicly responded to the allegations. The Trump administration’s stance aligns with U.S. tech companies like Anthropic, which has accused Chinese firms of improperly accessing its AI models. However, the Chinese government has not commented on the matter, leaving the dispute unresolved.
The upcoming talks are expected to focus on defining frontier AI models and establishing basic terms for cooperation. However, analysts suggest that significant progress is unlikely in a single meeting. The dialogue underscores the broader geopolitical tensions as both nations navigate the ethical, security, and economic implications of AI advancement.