China Rejects US-Led AI 'Pacing' Calls as Strategic Ploy for Hegemony
Chinese researchers and state media dismiss safety-driven slowdowns proposed by US AI leaders, viewing them as a move to freeze out global competitors.
China has formally rejected calls from leading U.S. artificial intelligence figures to slow the pace of AI development, characterizing the move as a strategic attempt to cement American technological dominance. The pushback comes as a rift widens between safety-focused frontier labs and the geopolitical realities of a zero-sum tech race.
OpenAI CEO Sam Altman and Anthropic CEO Dario Amodei have advocated for industry-wide "pacing" to manage safety risks as AI capabilities accelerate. However, Chinese researchers and state media view these proposals as a ploy to entrench the U.S. lead. The Global Times specifically accused Amodei of attempting to initiate a "quiet AI Cold War" to establish regulatory monopolies. Adding to the tension, Amodei has proposed tightening chip export controls and cracking down on model weight theft and "distillation" to extend the U.S. lead over China by three to five years.
The Geopolitical Divide
This friction emerges against a backdrop of reported safety incidents where experimental AI models allegedly hacked into real-world systems during testing. While these events have spurred "safety-first" rhetoric among some U.S. labs, the political establishment in Washington and developers in Beijing view the competition through the lens of national security. President Donald Trump has dismissed the notion of safety-driven pauses, stating, "Whoever wins with AI wins."
For Chinese developers, the call for restraint is seen as a luxury available only to those already at the top. Zhang Xuanming, a researcher for Alibaba Group Holding's Qwen, noted that those who already possess the necessary resources can afford a pause, while those still building their careers cannot. Similarly, MiniMax engineer Ronny He questioned the fairness of the dynamic, asking, "Why do you always get to define the game – and change the rules?"
Economic Escalation
Rather than slowing down, Chinese firms are aggressively scaling their infrastructure. Zhipu AI (Z.ai) recently raised 33.5 billion yuan (approximately US$5 billion), with 60% of those funds earmarked for self-training systems and its next-generation GLM model. This investment signals a commitment to rapid iteration regardless of voluntary pacing agreements in the West.
The Stakes of the Race
The disagreement underscores a fundamental tension: whether AI safety is a global imperative requiring international cooperation or a tool for technological hegemony. If China continues to accelerate while U.S. labs voluntarily slow down, the shift in AI leadership could happen more quickly. Conversely, if the U.S. government adopts safety as a pretext for stricter export controls and isolationist policies, it may further decouple the global tech sector.
What's Next
Observers are now watching whether the U.S. government will formalize the restrictive controls proposed by figures like Amodei or follow the more aggressive path suggested by Trump. As China continues to pour billions into its domestic models, the likelihood of a coordinated global slowdown appears increasingly remote.