Bloomberg Surveillance · Thursday, September 24, 2026
An expert emphasizes the need to build infrastructure for AI risk coordination early, drawing parallels to pandemic preparedness. The lack of a permanent channel for discussing AI risks between the U.S. and China is highlighted as a concern, especially in the event of a cross-border crisis.
“I think that you need to build the infrastructure for coordination early. You need to have the channels and have some idea of what information you're going to share at what time and build in domestic guardrails.”
“Right now, there is no permanent channel to discuss these AI risks. And if there were a cross-border crisis, like imagine hypothetically that we saw with OpenAI, Hugging Face was actually a U.S. company whose agents hacked into a Chinese company or vice versa. That becomes a pretty... scary cross-border crisis.”
“I think that this is just the most pragmatic way to actually make progress on this key issue.”