THE LONG VIEW

Bottom Line Up Front: The United States faces a critical strategic threat from China’s rapid pursuit of AI parity, which undermines prospects for genuine AI safety cooperation; the only viable path to future agreement is through a 'maximum pressure' campaign that expands the U.S. technological lead and forces Beijing into compliance. Threat Identification: China is actively working to close the AI capability gap with the United States—currently estimated at just eight months—through both legitimate development and illicit means including chip smuggling, distillation attacks on U.S. models, and exploitation of export control loopholes [Council on Foreign Relations, 2026]. Beijing views AI safety dialogues not as mechanisms for risk mitigation but as tools to gain access to restricted U.S. technology, undermining their integrity. Probability Assessment: Without intervention, China is highly likely to close or significantly narrow the AI gap within 12–18 months due to exponential progress and continued access to U.S. computing power via loopholes. The probability of China agreeing to binding AI safety constraints remains extremely low unless structural incentives change, such as a widened U.S. lead or a catalyzing crisis event akin to the Cuban Missile Crisis [Council on Foreign Relations, 2026]. Impact Analysis: If China achieves AI parity or superiority, it could deploy advanced AI for cyber warfare, intelligence operations, and military applications that directly threaten U.S. national security. Economically, Chinese tech firms would capture disproportionate global market share, eroding U.S. competitiveness. Moreover, unchecked AI proliferation increases the risk of non-state actors accessing dangerous capabilities, though Beijing prioritizes capability over collective safety. Recommended Actions: 1) Implement a 'maximum pressure' campaign by tightening AI-related export controls to eliminate all loopholes, including cloud-based access and third-country smuggling. 2) Establish a narrowly scoped U.S.-China AI safety dialogue strictly limited to technical risk mitigation, excluding discussions of export controls. 3) Accelerate domestic AI innovation and semiconductor production to extend the U.S. lead from eight to 18–24 months—an irreversible advantage in AI timelines. 4) Prepare for future negotiations from a position of strength, leveraging superior capabilities to compel Chinese compliance. Confidence Matrix: Threat Identification – High confidence (based on observed behavior and expert testimony). Probability Assessment – Medium-high confidence (extrapolated from current trends and historical analogs). Impact Analysis – High confidence (consistent with bipartisan national security assessments). Recommended Actions – Medium confidence (dependent on interagency coordination and global enforcement).