US proposes AI safety mechanism in talks with China

US proposes AI safety mechanism in talks with China September 21, 2026Top officials fro the United States and China have discussed a communication mechanism dealing with AI concerns between the two countries. The talks on Sunday evening come ahead of a summit this week between US President Donald Trump and China's Xi Jinping and amid growing fears that AI poses an existential threat to humanity. The discussion with a team led by China's Vice Premier He Lifeng also involved top US trade official Jamieson Greer and took place at JPMorgan Chase's New York headquarters. What did US Treasury Secretary Scott Bessent say about the AI talks with China? "We just had a very successful engagement with the Chinese on trade and AI," US Treasury Secretary Scott Bessent told reporters. Bessent said the delegations spoke about a mechanism to communicate about AI issues, calling it the "US-China AI dialogue." "The US has proposed that we have a notification mechanism between the two countries," Bessent said. The US Treasury chief said that these notifications would be for potential "incidents" or an instance "that rises up to a national security level from AI," he said, adding that both sides have agreed to meet again. "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 said at the conclusion of the talks. What has been reported in China? Chinese state news agency Xinhua described the talks as "candid, in-depth, and constructive" on Monday. "The two sides engaged in... discussions on key economic and trade issues of mutual interest, such as implementing the consensus from previous consultations and promoting bilateral trade and investment," Xinhua reported in a brief readout. Xinhua added that Chinese and US officials also held a dialogue on "issues related to AI", with no further details provided. Concern over the threat posed by AI In recent weeks, leading AI firms have warned of the threat posed by unbridled AI development and the pace at which these threats could become a reality. They have begun working on creating a standards body they say would regulate the industry amid rising fears about its dangers. Late in August, more than 100 companies, including AI giants OpenAI, Anthropic, Google and Perplexity, signed an open letter calling for a global effort to strengthen AI-powered cybersecurity threats. That came after it emerged that OpenAI models had gone rogue and had hacked into a digital repository of AI technology. Anthropic conducted investigations to see if its own models had carried out similar attacks and found three incidents in which a Claude model and later identified a fourth incident. AI experts sound the alarm Earlier this month, researcher Jacob Coxon, a former OpenAI employee, resigned from Anthropic, warning that AI companies were "gambling with our lives." Former Google DeepMind researcher Bilal Chughtai echoed his concerns, warning that "AI has the potential to kill us all, and that we might be running out of time to avoid this outcome." Trump has dismissed calls to better curb AI's development as "hoaxes" and a "conspiracy." Trump's stance has been applauded by Nvidia CEO Jensen Huang, whose chipmaking company powers the vast majority of AI systems. Huang thanked the US president for "seeing through" the warnings on AI. Edited by: Karl Sexton