In New York, the world's two dominant artificial intelligence powers sat down for eight hours and emerged with something rare: a shared acknowledgment that the race to build transformative technology carries risks neither nation can manage alone. US Treasury Secretary Scott Bessent and China's Vice Premier He Lifeng proposed a formal notification channel — a kind of hotline for the AI age — designed to surface national security threats before they spiral beyond reach. The meeting, held ahead of a Trump-Xi summit, did not resolve the deeper tensions over trade, chips, or critical minerals, but
US proposes AI safety notification system in high-level China talks
Moving from opaque to more transparency between the number one and number two AI powers
Mark
So they're proposing a notification system for AI threats. What does that actually mean in practice?
Mimi
It's a formal channel where if one side detects something that looks like a serious AI-related national security incident, they can alert the other. Think of it as a dedicated line for when things go wrong—a way to prevent miscalculation or escalation.
Luke
But we don't know what qualifies as "serious enough." That's going to be defined by each side separately, right? And there's no enforcement mechanism mentioned.
Mimi
True. It's more about creating a habit of communication than creating rules. The real value is that both sides are acknowledging they need to talk about this.
Mark
Why exclude chip export controls from this? That seems like the thing they'd actually want to negotiate.
Mimi
Bessent was signaling that this dialogue isn't about tightening restrictions. It's about managing risk through transparency, not through trade barriers.
Luke
But the US already has chip export controls in place. So by saying they're not part of this mechanism, he's essentially saying those controls stay as they are. It's not really a concession.
Mark
And the bigger trade truce expires in November. Does this AI agreement actually matter if the whole relationship blows up over tariffs?
Mimi
That's the real question. This is a small gesture ahead of the Trump-Xi summit, but it doesn't address the fundamental tensions—rare earths, tariff levels, the broader competition.
Luke
Right. And we don't know if Trump will even honor what Bessent negotiated. The summit hasn't happened yet.
Mark
So this is basically a confidence-building measure before the real negotiations begin.
Mimi
Exactly. A way to show both sides are willing to talk about the future, even if the present is still contested.
Le Pouls
- Two superpowers locked in a technological race agreed, at least in principle, that silence between them is more dangerous than dialogue.
- Eight hours of talks in New York produced a proposed AI notification system — but pointedly left export controls on advanced chips untouched, signaling Washington's limits on how far cooperation will stretch.
- A fragile trade truce struck in Busan is set to expire November 10, and neither side emerged from Sunday's meeting with a clear path to renewing it.
- Critical minerals — the rare-earth materials underpinning everything from electric vehicles to semiconductors — remain a live fault line, with no new movement announced.
- The discussions are landing as a carefully choreographed prelude to a presidential summit, with both sides projecting enough goodwill to keep talking while the larger contest between them stays unresolved.
In New York, the world's two dominant artificial intelligence powers sat down for eight hours and emerged with something rare: a shared acknowledgment that the race to build transformative technology carries risks neither nation can manage alone. US Treasury Secretary Scott Bessent and China's Vice Premier He Lifeng proposed a formal notification channel — a kind of hotline for the AI age — designed to surface national security threats before they spiral beyond reach. The meeting, held ahead of a Trump-Xi summit, did not resolve the deeper tensions over trade, chips, or critical minerals, but it marked a quiet admission that even rivals must sometimes speak plainly to one another.
Scott Bessent and He Lifeng spent eight hours in New York on Sunday laying groundwork for a presidential summit later in the week. What they produced was a proposal Washington called a breakthrough: a formal channel for the United States and China to notify each other when artificial intelligence incidents rise to the level of national security concern — a hotline, in effect, for the AI age.
Bessent framed the initiative as a matter of basic transparency between the world's two leading AI powers, while US Trade Representative Jamieson Greer made clear the dialogue would not extend to restricting advanced chip exports. The talks were built around shared vulnerability, not containment.
Beyond artificial intelligence, both sides discussed a framework for identifying goods that could move between the countries at reduced tariff rates — Chinese consumer products flowing one direction, American energy, agriculture, and medical devices the other. But the broader trade picture remained precarious. A truce reached in Busan had capped US tariffs at roughly 20 percent after a bruising escalation that briefly pushed duties into triple digits; that agreement expires November 10, and Sunday's meeting produced no resolution on the critical minerals — rare-earth magnets and the like — that both economies depend on.
Chinese state media described the exchanges as 'frank, in-depth, and constructive,' and a working group was set to continue talks the following day. The language suggested enough common purpose to sustain the conversation, even as the deeper contest between the two nations remained very much alive. The AI notification system was a small but telling concession to reality: when two rivals are racing to build technology that could reshape the world, they need at least one way to talk when something goes wrong.
Scott Bessent and He Lifeng sat down in New York on Sunday for eight hours of talks that would set the stage for a presidential summit later in the week. The US Treasury Secretary and China's Vice Premier emerged with what Washington called a breakthrough: a proposal for a new channel between the world's two dominant artificial intelligence powers, designed to flag threats serious enough to trigger national security alarms.
Bessent, flanked by US Trade Representative Jamieson Greer, framed the initiative as a matter of basic transparency. "Moving from opaque to more transparency between the number one and the number two AI powers in the world is very important," he said. The mechanism would create a formal way for both governments to communicate when they detected AI-related incidents that crossed into the realm of national security concern. It was, in essence, a hotline for the AI age—a recognition that two superpowers developing advanced artificial intelligence systems needed a way to talk when things went wrong.
What the proposal notably did not include was any restriction on the export of advanced AI chips or the manufacturing equipment used to produce them. Greer made that clear, signaling that Washington was not using this dialogue to tighten the technological screws it had already begun to turn. The talks were framed around shared vulnerability and common ground, not containment.
The meeting served as a dress rehearsal for the Trump-Xi summit scheduled for later that week in Washington. Both sides had work to do beyond artificial intelligence. They discussed a process agreed to in May for identifying goods that could move between the countries at lower tariff rates. Greer outlined the contours: China might export consumer goods and other low-tech products, while the US could send energy, agricultural goods, and medical devices. It was the kind of granular negotiation that happens in the margins of larger geopolitical contests.
But the broader trade picture remained fragile. A truce reached in Busan, South Korea, in October of the previous year had capped US tariffs on Chinese goods at roughly 20 percent, a settlement that came after a tit-for-tat escalation had briefly pushed duties into triple digits. That agreement was set to expire on November 10. Neither the AI dialogue nor the tariff discussions had resolved the question of rare-earth magnets and other critical minerals—materials essential to everything from automobiles to advanced semiconductors. Bessent and Greer announced no new movement on that front after Sunday's meeting.
Chinese state media characterized the discussions in measured terms, describing them as "frank, in-depth, and constructive exchanges" on economic and trade matters, including the artificial intelligence dialogue. A working group would continue talks on Monday, according to Bloomberg's reporting of Li Chenggang's comments. The language suggested both sides had found enough common purpose to keep talking, even if the larger contest between them remained unresolved. The AI notification system represented a small but significant acknowledgment: in a world where both nations were racing to develop transformative technology, they needed a way to manage the risks that came with it.
Citations marquantes
Moving from opaque to more transparency between the number one and number two AI powers in the world is very important— Scott Bessent, US Treasury Secretary
China and the US had frank, in-depth, and constructive exchanges on important economic and trade matters, including dialogue on issues related to artificial intelligence— Chinese state media characterization of the talks
Envie de l'histoire complète ?
Lire l'original sur Al Jazeera ↗
Politics & Governance