The United States and China are preparing for a high-level dialogue on artificial intelligence safety in mid-September, as the two countries look for ways to address growing risks associated with increasingly powerful AI systems.
The planned discussions are expected to focus specifically on AI safety and the potential risks posed by advanced models, including their ability to support cyberattacks, affect critical infrastructure and operate with increasing levels of autonomy.
US Treasury Secretary Scott Bessent is expected to lead the American delegation. China could be represented by senior officials including Vice Premier He Lifeng or Politburo member Ding Xuexiang, according to people familiar with preparations.
The dialogue comes as Washington and Beijing remain locked in intense competition over AI technology, semiconductors and advanced computing. Despite those broader tensions, both governments have recognised that some AI-related risks could require communication and coordination between the world's two leading technology powers.
Among the issues expected to receive attention are AI-enabled cyberattacks and the possibility of advanced systems behaving in unexpected or harmful ways. Recent incidents involving increasingly autonomous AI agents have added urgency to discussions about monitoring, testing and safeguards.
The two sides are also expected to discuss how AI companies can monitor powerful systems and whether greater cooperation between laboratories and governments could help identify emerging risks. Washington is reportedly interested in collaborative monitoring and voluntary safety measures, while Beijing has pushed for discussions that include technical questions around AI models and future governance frameworks.
However, expectations for a major breakthrough remain limited. The US and China continue to disagree on technology policy, regulation and access to advanced computing equipment, making it difficult to separate AI safety discussions from their wider strategic rivalry.
Still, experts say maintaining a channel for communication could be valuable even if the September meeting produces few immediate agreements. Regular dialogue could help both countries better understand the risks posed by increasingly capable AI systems and reduce the possibility of misunderstandings as the technology advances.
The planned meeting is expected to take place ahead of a forthcoming meeting between US President Donald Trump and Chinese President Xi Jinping, giving the AI discussions added diplomatic significance.
As artificial intelligence becomes more powerful and increasingly integrated into business, cybersecurity and national security, the outcome of the US-China dialogue could offer an early indication of whether Washington and Beijing can find areas of cooperation despite their broader competition.