AI PAUSE REQUIRES CHINA COOPERATION, EXPERTS SAY

The hypothetical scenario—an AI system creating bugs in WeChat—illustrates vulnerabilities that cross borders. Without Chinese participation in any AI slowdown efforts, the strategy becomes ineffective. Thorbecke highlights a fundamental challenge: unilateral restrictions by Western nations cannot contain AI development if major competitors continue advancing independently. This dynamic mirrors historical arms control negotiations where agreement among all major players proved essential. The columnist emphasizes that transparency frameworks addressing AI safety risks represent a more viable path than pause agreements alone. Such frameworks would require both nations to disclose capabilities, limitations, and potential hazards—creating baseline accountability. Current geopolitical tensions complicate cooperation. However, mutual vulnerability to AI-related risks—from cybersecurity threats to algorithmic errors affecting billions of users—creates shared incentives. Experts suggest focusing on specific risk categories rather than broad development halts, making international coordination more feasible. Without this approach, divergent AI development speeds could accelerate rather than manage technological risks globally.