US, China Weigh AI Incident Alert Channel

Two of the world's largest AI powers may be moving toward a channel for telling each other when artificial intelligence systems go wrong. Treasury Secretary Scott Bessent said the U.S. proposed a new “notification mechanism” for AI incidents that could affect national security, part of weekend discussions held ahead of talks at the White House this week between President Donald Trump and China's Xi Jinping.

Bessent described the idea to reporters on Sunday after meeting with Chinese Vice Premier He Lifeng in New York. The proposal was characterized as one element of the weekend discussions that also covered trade. A meeting between Trump and Xi is set to take place on Thursday, according to Bessent and U.S. Trade Representative Jamieson Greer.

Bessent Frames Shared Threats

Speaking after the New York talks, Bessent said the two sides should aim for a common understanding of the risks each faces. He said the U.S. proposed the notification mechanism and described transparency between the countries as important.

“We want a shared vision of common goals and common threats,” Bessent told reporters after talks with Chinese Vice Premier He Lifeng in New York. “We think that just like any cross-border activity, that moving from opaque to more transparency between the No. 1 and the No. 2 AI powers in the world is very important.”

— Scott Bessent, U.S. Treasury Secretary

The remarks came as Bessent and Greer stood together in the lobby of the JPMorgan Chase headquarters late Sunday. The description of the AI proposal was limited to the general concept of a notification mechanism, without detail on how it would operate.

The proposal, as described by Bessent, would cover AI incidents that could affect national security. He did not elaborate on what those incidents would include.

A Board of Trade, Operationalized

Alongside the AI discussion, Bessent and Greer said the groundwork had been laid for the Trump-Xi talks on Thursday. They said the two sides had agreed to operationalize a new “Board of Trade” that the two leaders had discussed when they met in Beijing in May.

“Today those notions have become reality,” Greer said.

— Jamieson Greer, U.S. Trade Representative

Greer also said the countries were still discussing which “nonsensitive” items could be considered for lower tariffs. He said the list could potentially include consumer and agricultural goods, energy products and medical devices.

That list remains under discussion, and Greer did not indicate when it might be finalized.

Trump's Stance on AI Speed

Trump has resisted calls to slow down AI development, saying that would help China catch up to U.S. companies. That position has been a consistent feature of his approach to AI policy, and it sits alongside the administration's proposal for a notification channel with Beijing.

The weekend discussions represent the latest round of engagement between the two governments on both trade and technology issues. The AI notification mechanism was described as a proposal, not an agreed arrangement.

Xinhua's Read on the Talks

Chinese state media offered a shorter account of the AI portion of the discussions. Xinhua said Bessent and the vice premier talked about issues “relating to AI” without mentioning specifics, and also had “candid, in-depth and constructive” discussions on key economic and trade issues.

The readout did not describe the notification mechanism or indicate how China views the U.S. proposal. Bessent's account came from the U.S. side.

Chen Sees a Precedent

One analyst said the significance lies in the fact that the two sides agreed to keep talking. George Chen, a partner and chair of digital practice at The Asia Group consultancy, offered written comments on the outcome.

“The fact that both sides agreed to continue the dialogue is significant,” said George Chen, a partner and chair of digital practice at The Asia Group consultancy, in written comments. “The initial outcome — an AI risk notification mechanism — sets a precedent that other countries may follow.”

— George Chen, partner and chair of digital practice at The Asia Group consultancy

Chen's comments point to the possibility that a bilateral arrangement between Washington and Beijing could become a template for other governments. That is his assessment, not a stated outcome of the talks.

What the Proposal Does Not Say

As described publicly, the notification mechanism comes with few operational details. The source material does not specify what types of incidents would trigger an alert, what information would be shared, or which agencies would operate the channel.

It is also not clear whether the mechanism would be voluntary, how quickly alerts would be shared, or whether either side would be obligated to disclose incidents. None of those points appear in the public statements from the weekend talks.

The proposal was described by Bessent, not released as a formal document or joint statement. The AI portion of the weekend discussions, as reported, consisted of the U.S. proposal and the two sides' agreement to continue the dialogue.

Trade Track Runs in Parallel

The AI notification mechanism was discussed alongside trade matters, including the Board of Trade and the potential tariff list. Greer's remarks on “nonsensitive” items identified consumer and agricultural goods, energy products and medical devices as possible candidates for lower tariffs.

Both Bessent and Greer linked the weekend discussions to the Thursday meeting between Trump and Xi. They said the groundwork had been laid for those talks and that the Board of Trade would be operationalized.

What the leaders will address on Thursday, beyond the items described, has not been detailed in the source material.

What to Watch After Thursday

For technology firms whose systems operate in both the U.S. and China, the proposed AI notification mechanism is the item with the most direct relevance. But as it stands, the proposal is a concept described by a U.S. official, without definitions, timelines, or a named authority to run it. Whether it becomes a working channel or remains a diplomatic phrase may depend on follow-through after Thursday's meeting, and on whether other governments treat it as a precedent, as Chen suggested they might.

Companies in the sectors Greer named — consumer and agricultural goods, energy products, medical devices — have a more immediate stake in the tariff list. For security leaders, the practical questions around the AI mechanism are still open: what counts as an incident, who reports it, and how quickly. Until those questions are answered, the mechanism remains a proposal. This suggests that the information governments may eventually seek from each other is information that, in practice, originates inside the companies building and operating AI systems — which could mean new expectations around how those companies detect, classify, and disclose AI incidents in the first place.

Related reading from SecurityWeek: What the Hugging Face Incident Teaches Security Leaders About AI Agent Access, Industry Reactions to OpenAI Models Hacking Hugging Face, and OpenAI Agents Coordinated via Makeshift Message Board Ahead of Hugging Face Hack.