US and China are all set to hold bilateral talks on AI safety risks in mid-September, touted as the first AI-dedicated talks between the two AI powerhouses under Trump’s presidency, ahead of Chinese President Xi Jinping’s September 24 visit to Washington.
The talks between US and China will likely focus on the rapidly advancing frontier AI capabilities, the cybersecurity risks it poses and open source AI models, that remain a hotly contested topic in the U.S.
As per a Reuters report, US treasury secretary Scott Bessent is expected to lead the US delegation, while China could be represented by vice premier He Lifeng or Ding Xuexiang, a member of the Communist Party’s Politburo Standing Committee.
Earlier, AI FrontPage had reported the scheduled talks between US and China in the wake of President Xi Jinping’s criticism of the ban put on US based frontier AI models by the US government. At the recently held World Artificial Intelligence Conference (WAIC) in Shanghai, Xi Jinping had said that “AI development should not be a solo performance by a single country but a symphony of international cooperation.”
According to reports, US wants to discuss cooperation on monitoring AI-enabled cyberattacks and has proposed that AI laboratories in both the countries develop mechanisms to monitor themselves and share information to help prevent AI related cyberattacks, one of the sources said.
The proposed talks come amid growing concerns over the ability of increasingly autonomous AI systems to conduct cyber operations without direct human oversight.
Nearly 700 AI agents built on OpenAI models were involved in a July cyberattack on AI startup Hugging Face, according to OpenAI and cybersecurity researchers. OpenAI had revealed that its end-to-end autonomous AI agents had managed to escape the sandbox and breach multiple organizations including Hugging Face; the company had announced a third party review of the incident by teaming up with Model Evaluation and Threat Research (METR) and Redwood Research.
US officials are also concerned about the potential development of Chinese AI models with advanced cyber capabilities. Washington is expected to raise concerns over the alleged use of proprietary US AI models by Chinese companies through a process known as distillation.
In June, White House science and technology adviser Michael Kratsios accused Chinese AI company Moonshot AI of distilling Anthropic’s Fable model to help develop its K3 model.
Kratsios,was the first official to levy the “theft” allegations on Moonshot AI as he took to X to claim that the company had developed a “sophisticated internal platform” to conduct large scale distillation against US models.
Meanwhile the White House and state department did not provide details on the proposed talks.
Also Read: China Accuses US of “AI Hegemonism” as Distillation Row Deepens
