US-China AI Risks Common Ground

• Safety guidelines framework: Proposed joint US-China protocols would establish output guardrails for cyber, chemical, and biological AI applications, identifying high-risk use cases and limiting malicious capabilities across systems[1] • Supply chain control: The US maintains semiconductor advantage through advanced chip export restrictions (though some controls have been eased), while China invests billions in indigenous chip production to narrow this gap[3] • Infrastructure asymmetry: China's manufacturing dominance, energy surplus, and centralized coordination create asymmetric advantages in deploying AI infrastructure at scale that Western competitors would struggle to match[3] • Intergovernmental dialogue mechanisms: First US-China AI dialogue held in Geneva (May 2024); 2026 anticipated for institutionalized dialogue specifically focused on non-state actor risks and high-risk applications[1]

The geopolitical fragmentation evident in reduced US-China engagement at military AI forums (REAIM 2026 vs. 2024) threatens to create a patchwork of inconsistent AI governance policies globally. If great powers remain aloof from cooperation, middle powers may drive forward independent rules of the road, potentially creating competing standards. The convergence of US-China interests on non-state actor risks presents a rare alignment opportunity, but success depends on institutionalizing dialogue before domestic political transitions or strategic competition further diverge policy approaches. Failure to establish joint safety guidelines risks accelerating the 'safety arbitrage' problem where malicious actors exploit jurisdictional differences.