The US and China are preparing to hold high-level talks on AI in September, a sign that both governments see continued engagement as necessary despite an intense strategic rivalry.
The meeting stems from agreements reached during the Trump-Xi summit in May and is expected to take place before Chinese President Xi Jinping’s planned Sept. 24 visit to the US. While neither government has formally announced a date, officials on both sides are continuing preparations, with the agenda, location and participating agencies still under discussion.
US Treasury Secretary Scott Bessent is expected to lead the American delegation. Other participants have not been finalized. This proposed meeting is the first formal AI dialogue between the two countries under President Trump.
An Issue Beyond Commercial Technology
The discussions demonstrate a growing recognition in both capitals that AI has become a strategic issue extending well beyond commercial technology. Washington and Beijing are weighing policies governing advanced AI systems capable of enabling cyberattacks against critical infrastructure and reshaping labor markets. Both governments are also evaluating how to manage the rapid development of frontier AI models while limiting security risks.
Although export controls on advanced semiconductors remain a major source of tension, AI governance is expected to dominate the September discussions. US officials have sought safeguards around the release of powerful AI models and measures designed to prevent sophisticated systems from reaching malicious non-state actors.
Earlier this year, Bessent said future talks should focus on limiting the proliferation of advanced AI technologies while establishing common expectations for their deployment.
Chinese officials are expected to raise concerns about Anthropic’s Mythos and the future release of advanced US frontier models.
Intellectual property is expected to feature prominently. In a recent interview, Bessent said US officials had identified evidence suggesting American LLMs had been incorporated into Chinese systems without authorization, adding that Washington intends to examine the issue further.
China is expected to arrive with its own priorities. Analysts say Beijing wants detailed discussions about how the US plans to oversee future advanced AI releases, whether additional restrictions could be placed on Chinese open-weight models and how Washington intends to approach safety standards for leading AI systems.
No Breakthrough Expected
The talks will occur during a period of heated AI competition between the two superpowers. The US has already imposed restrictions on exports of advanced AI chips to China, while weighing additional measures affecting Chinese AI developers. Beijing has reportedly considered limiting overseas access to some of its own most capable AI models because of security concerns.
Despite those disputes, both governments appear to recognize that maintaining communication reduces the risk of misunderstanding as AI capabilities advance. Previous exchanges under the Biden administration produced agreement that humans should retain control over decisions involving nuclear weapons, but AI dialogue stalled after initial meetings.
Analysts caution that the meeting is unlikely to produce major policy breakthroughs. Instead, the first objective may be establishing common terminology and identifying areas where technical discussions can continue even as US-China competition intensifies.
