US officials see little room for progress in AI safety negotiations with China before President Donald Trump’s meeting with Chinese President Xi Jinping, according to WIRED’s report on administration officials’ assessment. The talks have occurred at a technical level, but a proposed incident hotline between Treasury Secretary Scott Bessent and Chinese Vice Premier He Lifeng is still not operational, the report says.

The immediate security-policy development is not a new bilateral AI agreement. It is the Trump administration’s move toward a rule restricting how Chinese firms can access US-made advanced chips through overseas data centers. WIRED reports that the administration announced this week that it would move forward with such a rule, while leaving open whether Trump would raise Chinese interference and related AI concerns directly with Xi.

The proposed measure would target remote access rather than only physical shipments into China. In earlier reporting on the administration’s deliberations, WIRED said Trump chose an interim trade rule aimed at preventing Chinese AI laboratories from leasing computing capacity from data centers in Southeast Asia, instead of immediately placing Megaspeed on the US entity list. The report said Megaspeed subsidiaries had hosted servers in Malaysia and Indonesia and rented computing power to Chinese companies.

That distinction defines the current scope of the change: Washington is pursuing a mechanism to limit overseas use of US chips while keeping some technology commerce with China possible. It is not evidence that the two governments have agreed on shared AI-incident reporting rules.

The negotiations also face a basic disagreement over priorities, according to WIRED. US officials are focused on autonomous or misaligned AI systems, while Chinese officials are described as more concerned about open-weight models being used by domestic critics and dissidents. WIRED also reports that China has rejected allegations involving Alice Huang and Chinese intelligence as “groundless” and a “smear.”

The evidence therefore establishes an advancing US technology-control measure alongside stalled diplomatic coordination. It does not establish that the proposed rule is fully implemented, that the hotline is operating, or that Trump and Xi will reach an AI agreement at their meeting.