Chinese leadership might not be convinced by AI extinction risks. However, they might agree to a slowdown deal if they realize the US compute advantage is insurmountable in the short term. The deal offers them a chance to catch up on algorithms via transparency, avoiding a default loss.

Beyond official diplomacy, direct meetings with US leaders like President Trump provide unfiltered data for Beijing. China's intelligence services analyze off-the-cuff remarks about domestic politics and strategic priorities to build a psychological profile and assess US weaknesses.

The US and USSR, despite being adversaries, collaborated to prevent nuclear proliferation to rogue actors. A similar model can be applied to AI. The US and China share an interest in preventing powerful open-weight models from being used for cyber-attacks or bio-terrorism by third parties, creating a foundation for a safety dialogue.

The most realistic path for US-China AI talks involves creating protocols to prevent models falling into the hands of terrorists or criminal groups. Cooperation on preventing state-level cyberattacks is unlikely, as both sides assume the other's intelligence agencies will use AI for espionage.

Emerging technologies like confidential computing offer a novel path to building trust between rivals. They could allow for mutual verification of AI systems or even sensitive government documents, with an AI providing trusted attestations (e.g., "no backdoor found") without exposing the underlying secret data.

Contrary to the belief that China is enigmatic, its strategic ambitions are openly stated. The primary barrier to understanding their intentions is not sophisticated encryption or secrecy, but the linguistic and cultural challenge of interpreting their public documents and statements in their original context.

The fact that both US and Chinese AI labs are predominantly building on the same transformer-based architecture is a major geopolitical advantage. It means both sides encounter similar unexpected behaviors and safety issues, creating a shared understanding that can be a basis for cooperation.

Global consensus on AI safety is unlikely soon. The most practical approach is a bilateral agreement between the US and China, the two dominant players. This can begin with informal 'track two' conversations between their respective AI labs before escalating to a formal government pact.

Despite sounding significant, the new "China-U.S. AI Dialogue" is merely a mechanism for countries to notify each other of national security-level AI incidents. U.S. Treasury Secretary Scott Besant's comments reveal it is not an agreement to halt or slow down AI development, which aligns with the Trump administration's public stance.