Published Updated

Technology and Science20 September, 20263 min read

Trump And Xi Summit: U.S. Proposes AI Safety Notification Mechanism For Incidents

The US proposed a cross-border AI incident notification mechanism to China ahead of Trump-Xi summit Trump announced creation of AI Force and AI czar, rejecting calls to slow AI development

Read the story
Trump And Xi Summit: U.S. Proposes AI Safety Notification Mechanism For Incidents
Image: 조선일보

Trump and Xi set AI agenda

U.S. President Donald Trump is set to host Chinese President Xi Jinping at a summit in Washington, where technology, artificial intelligence, semiconductors, and rare earth metals are described as central to strategic competition between Washington and Beijing.

U.S. Treasury Secretary Scott Bessent and Chinese Vice Premier He Lifeng concluded talks in New York on Sunday, during which the American side proposed a new AI safety notification mechanism for Trump and Xi to consider at their summit this week.

Image from 104.5 WOKV
Image: 104.5 WOKV

U.S. and Chinese officials are also working to finalize a delegation of top Chinese business leaders accompanying Xi, with possible participation of executives from major companies including BYD and Xiaomi, and CATL.

The Al-Bayan account frames the AI file as extending beyond model performance to advanced chips, data centers, energy, cloud computing, supply chains, regulatory standards, and each country’s ability to build an integrated technological ecosystem.

The same account says the summit may seek a narrow space for cooperation, such as opening channels of communication about AI-related cyber incidents or setting general safety principles, without implying an end to the race for technological supremacy.

Reporting for this sectionAl-Bayan

Guardrails debate and “AI Force”

U.S. President Donald Trump rejected calls for an AI slowdown and instead announced he would establish an “AI Force” and appoint a special envoy, while dismissing safety concerns as a “hoax.”

Trump posted on Truth Social that “AI taking over the World, destroying Humanity, and all other things bad, is a HOAX.”

Image from 14 News
Image: 14 News

Trump also promised that “we will also be looking for BAD, and we can do that, very easily, with our already existing Criminal and Civil Justice System,” as he described the initiative as a way to watch over the industry as it grows.

The Hill described Trump’s response to rising demands for guardrails as he compared the “AI Force” to the U.S. Space Force created during his first term, while providing few details about its makeup or responsibilities.

The Hill also reported that Treasury Secretary Scott Bessant laid groundwork for more global cooperation on AI in talks with Chinese counterparts in New York on Sunday, including a mechanism for alerting each other of potential national security issues related to AI.

Reporting for this sectionDWThe Hill

Notification system and stakes

The United States proposed a notification mechanism for AI incidents that could pose a threat to “national security,” and the proposal was raised during negotiations in New York between U.S. Treasury Secretary Scott Bessent and Chinese Vice Premier He Lifeng.

Bessent said, “We want a shared vision of common objectives and threats,” after meeting with He Lifeng, and he described the effort as moving from opacity to greater transparency in a strategic area.

The proposal’s details remained “unknown,” and the initiative also had not been confirmed as accepted by China, according to the Perfil account.

The stakes extend to how AI governance is handled across borders, because the Ars Technica framing ties Trump’s approach to a defense of the emerging industry while industry figures call for meaningful regulation and oversight.

The CBC account adds that AI guardrails are being demanded as policymakers struggle to keep up with an “increasingly volatile artificial intelligence race,” and it quotes Dario Amodei arguing that “the incentives to pull ahead and the military advantage that you get from that” make verification “ironclad.”

Reporting for this sectionPerfilCBC