This article critically analyzes the dangerous US-China AI "arms race," emphasizing significant, shared risks like autonomous agent misbehavior. It argues that China's advanced AI development invalidates US regulatory inaction excuses, despite China's own rules. The author rejects the "inevitability" narrative, stating current government-led risk mitigation is insufficient. Instead, direct information exchange between companies and researchers is urged to prevent escalating incidents. Policymakers must act decisively, shifting from competitive development to cooperative prevention before unchecked AI leads to severe global consequences.