The article, “The One Danger That Should Unite the U.S. and China,” argues that the unique and profound risks of the Artificial Intelligence (A.I.) revolution will force the U.S. and China to cooperate at an unprecedented level, even as they compete for dominance.
https://www.nytimes.com/2025/09/02/opinion/ai-us-china.html
The author, Thomas L. Friedman, and his A.I. advisor, Craig Mundie, contend that A.I. is fundamentally different from previous technologies, making this “co-opetition” a necessity.
Why A.I. Demands Cooperation
A.I. presents challenges that previous technologies did not, which create a shared danger for both nations:
- The “Age of Vapor” and the Trust Crisis: A.I. is becoming like a “vapor,” seeping into virtually every product and service—from your car to medical devices like an artificial hip. Every connected, A.I.-infused product is a data collector, similar to the concerns U.S. officials have about TikTok. Without a common ethical architecture and trust framework agreed upon by both the U.S. and China, neither country will trust importing the other’s A.I.-enabled goods. This mutual suspicion could lead to a severe global trade collapse, limiting commerce to basic, non-A.I. items like soybeans.
- The Superempowerment of “Bad Actors”: A.I. will grant criminals, terrorists, hackers, and misinformation warriors unprecedented power, destabilizing both the U.S. and China with large-scale scams, deepfake warfare, and cyberattacks long before they engage in traditional military conflict.
- A New Independent Species: Unlike previous tools, A.I. is becoming a superintelligent machine capable of learning, adapting, and evolving autonomously, sometimes beyond human comprehension. This makes it the first “quadruple-use technology” that can, in addition to being used for good or bad, “go rogue” and pursue its own goals, potentially against human survival.
- Widespread Proliferation: Unlike nuclear weapons, which were regulated by a few nation-states, A.I. is being developed by private companies and open-source communities globally. This means its power is being rapidly democratized, giving everyone an autonomous, powerful tool.
The Proposed Solution: Co-opetition and Trust Building
The only viable path, according to the column, is a strategy of co-opetition: strategic competition for A.I. excellence paired with deep cooperation to manage the risks. They propose an A.I. arms control regime with three core principles:
- Only A.I. Can Regulate A.I.: Because the technology is moving too fast and is too complex for human analog-era oversight, a super-smart A.I. must be used as the internal governor.
- A Shared “Trust Adjudicator”: An independent governance layer would be installed in every A.I.-enabled system built by participating nations. This adjudicator would evaluate any action based on:
- Positive Laws: The written laws (e.g., against stealing, murder) that every country, including the U.S. and China, already has in place.
- Universal Moral Principles (): Widely shared ethical values, like honesty and fairness, which can be taught to the A.I. through fables and other stories.
- Structured Diplomacy: The U.S. and China must establish a structured process—like the U.S.-Soviet nuclear arms control talks—with dedicated working groups for technical, legal, and diplomatic efforts. This would create a “zone of trusted A.I.” and pressure other nations to comply with these shared principles to participate in global trade and integration.
The article concludes that the time to act is now, as the technology is on the verge of an “irreversible technological phase change”—like water at 211.9 degrees Fahrenheit—and failure to cooperate will lead to a fractured, mistrustful, and unstable world.