The author says this week’s fights—U.S. accusations of Chinese model “distillation” and OpenAI’s call for mandatory safety rules—show the same gap: models update on a product calendar while governments still move on a legislative one.

Governance is treated as part of the export stack. Buying U.S., Chinese, or EU AI also imports who logs incidents, who can inspect or pause a system, and whose standards travel. Washington sells a full stack; Beijing a Global South standards club; Europe a slow, legal market-access model.

Proposed fix: aviation-style two clocks—parliaments set durable red lines; regulators issue narrow, expiring directives from shared incident reports and pooled tests—so middle powers can demand audit and rollback rights instead of becoming rule-takers.

Washington is exporting an AI technology stack, Beijing is building a governance coalition, and Europe is revising its timetable. Strategic advantage will belong to states that can turn new evidence into legitimate rules before the next model generation arrives.

Within forty-eight hours, the AI race exposed two of its least settled rules. On September 8, Washington accused six Chinese developers of using distillation to extract capabilities from American models; Beijing rejected the charge as an attempt to suppress competition. On September 9, OpenAI said that voluntary commitments were no longer sufficient for increasingly capable systems and called for mandatory, capability-based national rules covering testing, independent assessment, cybersecurity and incident reporting.

One dispute concerns how models learn from rivals; the other concerns how governments learn from models. Together they reveal a strategic constraint. Developers can alter capabilities on a release calendar, while public institutions still respond on a legislative one. Industry proposals warrant scrutiny because incumbents can shape rules to their advantage, but the underlying mismatch is difficult to dismiss. A country may lead in chips, models and data centers yet surrender strategic initiative if it cannot convert new capabilities and new failures into credible rules before the next generation arrives.

Washington already describes artificial intelligence as a geopolitical contest. The official US AI strategy says that the country with the largest ecosystem will set global standards and collect economic and security benefits. A White House order promoting full-stack American AI exports packages hardware, cloud services, data systems, models, cybersecurity and applications for allies and partners and explicitly seeks worldwide adoption of American standards and governance models.

Beijing is building a competing route to influence. At July’s World Artificial Intelligence Conference in Shanghai, participants from more than one hundred countries and international organizations endorsed “agile governance,” faster revision of standards and their mutual recognition. The meeting also marked the signing of an agreement to establish the World Artificial Intelligence Cooperation........

© Eurasia Review