US, China eye AI dialogue as competition over frontier models intensifies

China and the United States are expected to hold talks on artificial intelligence in September to discuss the growing risks and policy challenges posed by powerful frontier AI models, according to Reuters. The dates for the meeting have not yet been finalised.

The two countries have grown increasingly wary of each other's AI capabilities in recent years, and the dialogue is expected to ease tensions while opening a channel for discussions on AI governance. The US delegation will be led by Treasury Secretary Scott Bessent, Reuters reported.

First US-China AI talks under Trump

The planned meetings will mark the first official US-China dialogue on artificial intelligence under President Donald Trump. The two countries last held AI discussions in 2024 under then-President Joe Biden, when both sides agreed that decisions on the use of nuclear weapons should remain under human control.

The US has already restricted exports of some of its most advanced AI chips to China. Reports also suggest Washington is considering limits on the use of Chinese open-weight AI models by US companies as their capabilities continue to advance.

More from Tech

China, meanwhile, is weighing restrictions on overseas access to its most advanced homegrown AI models amid concerns over the hacking capabilities demonstrated by Anthropic's unreleased Mythos model.

Bessent had said in May that future discussions would focus on preventing powerful AI models from reaching non-state actors. Following Trump's visit to China, Beijing confirmed that both countries had agreed to establish an intergovernmental AI dialogue, although it has not provided further details since.

China seeks technical AI engagement

According to analysts, Beijing plans to raise several key issues during the talks, including Anthropic's Mythos model, how the US intends to govern the release of future frontier AI systems, and whether Washington plans to impose restrictions on Chinese open-weight AI models.

The Trump administration is also discussing voluntary AI safety standards with major US AI companies, though it has yet to announce a formal regulatory framework.

Chinese officials are expected to spend the coming weeks finalising their position and deciding which ministries will participate in the dialogue. Sources said Beijing is seeking substantive technical engagement on AI governance rather than broader political negotiations.

Analysts expect the inaugural meeting to focus on establishing common definitions and basic principles for frontier AI, laying the groundwork for future discussions instead of attempting to resolve major policy disagreements.