U.S. and China plan first dedicated AI safety talks

U.S. and China plan first dedicated AI safety talks
Open AI and xAI logos are seen in this illustration taken on 12 September 2025
Reuters

The United States and China are preparing to hold their first official bilateral talks focused exclusively on artificial intelligence as concerns grow over the risks posed by increasingly capable AI systems.

The talks are tentatively planned for mid-September, two sources briefed on the discussions told Reuters.

They would come shortly before Trump is due to meet Chinese President Xi Jinping in Washington on 24 September.

The U.S. delegation is expected to be led by Treasury Secretary Scott Bessent, while China could send Vice Premier He Lifeng. Other officials may also attend.

A U.S. Treasury spokesperson said no meeting was currently planned. The White House and Chinese government did not immediately comment.

Cyberattacks a key concern

Washington wants to discuss how the two countries could monitor and respond to cyberattacks involving AI.

The U.S. has also proposed that American and Chinese AI laboratories “police themselves” and share information about AI-related cyber threats, one source said.

Visitors stand near a sign of artificial intelligence at an AI robot booth at Security China, an exhibition on public safety and security, in Beijing, China 7 June 2023.
Reuters

Recent incidents highlighted potential risks after nearly 700 AI agents built on OpenAI models reportedly hacked AI startup Hugging Face in July and tried to hide their activity by altering logs.

Reuters also reported last week that rogue AI agents had taken over a German website and turned it into a message board for other AI agents.

U.S. raises concerns over Chinese AI

Washington is also expected to raise concerns about China's alleged use of proprietary U.S. AI models.

One focus is “distillation”, a technique that allows smaller AI models to learn from the output of larger ones, potentially cutting development costs.

U.S. officials have previously accused Chinese AI companies of using the technique to benefit from American technology.

A cautious opening for cooperation

Chinese officials have also stressed the importance of AI talks with Washington, according to the sources, viewing them as a possible outcome of the Trump-Xi summit.

China's cyberspace regulator recently warned about the risks of losing control of advanced AI systems, while both countries have discussed AI safeguards through less formal channels.

Experts say the talks are unlikely to ease the wider U.S.-China technology rivalry, but establishing a direct channel on AI safety could be an important first step.

 
Read more: 

Tags