Artificial-intelligence safety moved onto the agenda of high-level U.S.-China economic talks in New York on Sunday, alongside trade and critical minerals. Reuters reported that U.S. Treasury Secretary Scott Bessent and Chinese Vice Premier He Lifeng discussed possible AI guardrails as the two governments prepare for a presidential summit.
The talks are preliminary, and AI remains entangled with the larger strategic competition over chips, data centers and technological leadership. But its appearance in an economic negotiation is notable: frontier-AI risk is beginning to sit alongside established bilateral issues rather than remaining a specialist technology-policy subject.
The hardest issues include systems capable of assisting with biological or nuclear weapons and the possibility of increasingly autonomous or self-improving AI. Any meaningful agreement would have to coexist with export controls and deep mistrust over each country’s intentions.
AI Safety Watch 