US-China establish AI incident notification channel… Bessent expects exchange of model testing practices

U.S. Treasury Secretary Scott Bessent said Washington and Beijing have begun a dialogue on managing shared AI risks and are pursuing a channel to notify each other of major AI incidents, speaking at a Council on U.S.-China Relations dinner in New York. Bessent said the two sides expect to exchange practices on pre-release evaluation and testing of advanced AI models and on preserving human judgment in AI deployment, while cautioning that shared agendas do not mean identical interests or approaches. The initiative matters because it creates a formal communication mechanism between the world's two largest AI markets, reducing the risk that a serious model failure or safety incident goes unexplained across the Pacific. What to watch next is whether the notification channel is formalized and whether the planned exchange of testing practices produces concrete, reciprocal commitments.