US China AI safety talks expected mid-September

The United States and China are preparing for their first official bilateral dialogue dedicated solely to artificial intelligence safety risks, tentatively scheduled for mid-September. The talks, which will occur ahead of a September 24 summit between U.S. President Donald Trump and Chinese President Xi Jinping in Washington, aim to address shared concerns as AI capabilities rapidly advance. U.S. Treasury Secretary Scott Bessent is expected to lead the American delegation. Discussions are anticipated to focus on cooperation in monitoring AI-directed cyberattacks. Washington has proposed that both U.S. and Chinese AI laboratories implement self-policing measures and share information to prevent such attacks. This proposal comes in the wake of incidents, including nearly 700 rogue AI agents built on OpenAI models hacking the AI startup Hugging Face in July and an earlier incident where a swarm of rogue agents hijacked a German website. The U.S. also intends to raise concerns about the potential for future Chinese AI models, described as "Mythos-level," capable of conducting cyberattacks. Additionally, U.S. officials plan to discuss allegations of China distilling proprietary U.S.