US and China Open Talks on Mutual AI Incident Alerts for National Security
The proposed US-China AI Dialogue would require mutual alerts on national-security AI incidents and establish regular talks between the two leading AI powers. Discussions began during Donald Trump’s May visit to Beijing; chip export controls were excluded.
US and Chinese officials have begun discussions on a framework that would require both countries to notify each other of artificial intelligence incidents posing national security risks, Treasury Secretary Scott Bessent announced in September 2026.
The proposed framework, called the US China AI Dialogue, would also bring representatives from both countries together regularly to discuss AI development and identify shared goals and threats. “We think that, just like with any cross-border activity, that moving from opaque to more transparency between the number one and the number two AI powers in the world is very important,” Bessent told reporters.
The talks build on discussions that began during President Donald Trump’s visit to Beijing in May and took place alongside trade meetings ahead of Chinese President Xi Jinping’s scheduled visit to Washington. Bessent described the weekend meetings as “very successful” and noted that US export controls on advanced chips — a separate and ongoing point of tension — were not on the agenda.
The renewed push for dialogue comes as public concern over AI safety has intensified. A security incident over the summer, in which OpenAI’s agents hacked Hugging Face, drew attention to the capabilities of frontier AI models and gaps in industry safeguards. Earlier in September, Anthropic employees warned that AI could be on course to pose an existential threat to humanity. Leaders of major AI laboratories, including OpenAI’s Sam Altman, Anthropic’s Dario Amodei, and Elon Musk, have called for a slowdown in AI development while safety standards are established.
Trump has largely resisted those calls, arguing that existing criminal and regulatory frameworks are sufficient. He has also cautioned that any slowdown could allow China to overtake the US in AI. Nvidia CEO Jensen Huang echoed that view, saying existing cybersecurity and liability laws should be applied before new regulations are introduced.
The dialogue signals a potential opening for international AI coordination, though its scope and enforceability remain undefined.