US-China AI Competition & Global Governance Gap

AI roleplay emerges as a massive but underreported use case—OpenRouter data shows it burned more tokens than coding. China's new regulation (July 15) is the first of its kind targeting roleplay, and labs like MiniMax and DeepSeek are actively embracing it, contrasting with US labs' caution, adding a cultural/regulatory dimension to US-China AI competition. A critical new piece highlights that Western AI discourse systematically excludes China, despite its massive role—the Stanford letter with zero Chinese signatories is a stark example. The two-level gap (model adoption and economic impact) adds nuance to US-China AI competition, reinforcing the need for neutral research bridges. California's AB 2392 could restrict GenAI in universities, adding a new regulatory angle. Xi's reaffirmation of open-source AI while tightening regulation signals China's long-term strategy: owning the global AI stack via open source, but with stability-first governance. The World AI Cooperation Organization is a new governance bid.