Why the U.S. Should Not Cooperate with China on AI Regulation
Source: Helen Raleigh. "Cooperating With China On AI Regulation Is A Stupid Idea." September 21, 2026. thefederalist.com
The Gist
The author argues that trying to team up with China on AI safety rules is a bad idea because China has broken its promises to the U.S. many times before—on trade, military bases, and drug control. Since China sees AI as a tool for its own power and control rather than a safety issue, the U.S. should focus on staying ahead in AI and setting its own rules instead of trusting China to cooperate.
Conclusion
The United States should not pursue AI regulation cooperation with Communist China, and should instead focus on maintaining AI leadership and self-governance.
Premises
- Historically, whenever the U.S. has sought cooperation from the CCP on major issues (WTO entry, South China Sea militarization, fentanyl export controls), the CCP has used these agreements to advance its own interests rather than honor commitments.
- China's WTO entry led to unfulfilled market-opening promises, intellectual property theft, forced technology transfer, and significant U.S. manufacturing job losses without the promised political liberalization.
- Xi assured Obama China would not militarize artificial islands in the South China Sea, but China did so extensively within three years.
- China made commitments to limit fentanyl precursor exports to both the Trump and Biden administrations, but enforcement has been inconsistent and China remains a primary source of drug-related chemicals.
- The CCP views AI as a strategic tool for military strength, social control, and competitive advantage over the U.S., not as a technology requiring safety cooperation for humanity's benefit.
- Chinese state media (Global Times) has explicitly rejected U.S. calls for AI cooperation as attempts to 'contain' China, and China's intelligence chief has framed AI safety concerns in terms of regime security rather than human safety.
- The CCP's handling of COVID-19 (suppressing information, blocking independent investigation) demonstrates a pattern of prioritizing party security over public/global safety when the two conflict.
Assumptions
- Past CCP behavior in unrelated domains (trade, territorial expansion, drug enforcement, pandemic response) reliably predicts its behavior in the AI domain.
- The CCP acts as a unified, strategic actor whose stated commitments are inherently untrustworthy rather than context-dependent.
- Any voluntary AI slowdown by the U.S. would be perceived and exploited by China as a strategic opportunity rather than reciprocated.
- U.S. self-governance and competition-focused strategy is sufficient to manage AI risks without any international coordination with China.
- The proponents of cooperation (Altman, Clinton) are naive or insufficiently informed about CCP behavior patterns.