Treasury Secretary Scott Bessent opened the door to talks on artificial intelligence (AI) risks with China, saying Washington will raise shared dangers at economic meetings in New York this weekend.

Bessent attached one limit. Washington will not slow its own AI development, and it still treats the technology as a race it intends to win.

AI Risk Open in New York Before the Trump-Xi Summit

Bessent and United States Trade Representative (USTR) Jamieson Greer will meet Chinese Vice Premier He Lifeng for several hours, Axios reported.

The agenda covers AI, trade, and rare earths. The meetings land eight days before President Donald Trump hosts Xi Jinping in Washington on September 24.

“The United States remains the leader in AI, and we are open to discussions on avoiding shared risks and avoiding bifurcation of our two systems. We expect our discussions to cover both open and closed weight models,” read the report, citing Bessent.

Bifurcation means the two countries drifting into separate AI systems that no longer work together. Open-weight models are ones whose underlying settings are published for anyone to download and reuse.

Both Governments Have Rejected Industry Calls to Slow AI Down

Anthropic chief executive Dario Amodei published an essay on September 12 urging rivals to slow capability gains. Trump rejected an AI slowdown two days later, calling it a hoax.

China’s foreign ministry called the slowdown “fear-mongering.” It said all parties should promote open and inclusive AI development.

BeInCrypto flagged that split on September 13, when Xi called for a consensus-based global governance framework while Washington pushed speed. A congressional push to pause frontier work has stalled.

Axios reported that a major safety breakthrough remains a tall order and that Bessent’s role signals engagement rather than a deal. Neither government has said it will limit its own development. September 24 is the test.

Source: BeInCrypto