
by Jianli Yang
The upcoming Donald Trump-Xi Jinping-Vladimir Putin summit will test whether great-power diplomacy can build a safer international order.
A trilateral summit alongside the APEC leaders’ meeting in Shenzhen, China would bring presidents Donald Trump, Xi Jinping and Vladimir Putin together at a moment of profound flux in the international order. The three leaders could discuss the Ukraine War, tensions over Taiwan, trade and strategic stability. But one question deserves particular attention: can the world’s three preeminent nuclear powers cooperate to prevent artificial intelligence from making nuclear war more likely
The United States increasingly questions the costs of maintaining the international order it built. China seeks to reshape that order while retaining its economic benefits. Russia challenges essential elements of it through military force. All three are dissatisfied with the existing arrangements, though for different reasons.
Could their strategic triangle provide a stable replacement
Henry Kissinger recognized that international stability requires more than a mechanical balance of power. Building on that insight, I propose a three-dimensional test: a sustainable equilibrium in the distribution of power, mutually accepted rules and a sufficiently legitimate foundation of values.
The triangle partly satisfies the first dimension. America remains the most comprehensive global power. China is its principal technological and economic competitor. Russia, though considerably weaker economically, remains a nuclear and energy superpower. Their capabilities make each indispensable to major strategic calculations.
But the triangle falls short on rules and values. The competing ambitions of the great powers offer little basis for mutually accepted rules. Moreover, agreements between three leaders also cannot command legitimacy if they all disregard the sovereignty and interests of other nations. Personal understandings are also vulnerable to changing administrations and strategic calculations.
The triangle may therefore produce useful bargains without creating a durable international order. Yet the intersection of artificial intelligence and nuclear weapons offers a meaningful test of whether the three powers can move beyond bargaining and toward building.
AI is transforming the foundations of national power, with the United States leading in frontier research, advanced computing and technological entrepreneurship. China combines rapidly advancing AI capabilities with extraordinary manufacturing capacity and increasingly affordable models distributed internationally. Russia cannot match either country’s advanced AI capabilities. Nevertheless, its energy resources, nuclear-energy expertise and military experience with drones and electronic warfare give it a consequential role.
What is emerging is an increasingly bipolar AI competition that intersects with a nuclear triangle and a more widely distributed geography of energy power. This combination creates serious dangers.
Nuclear deterrence has traditionally depended on reliable warning systems, secure communications and political leaders having sufficient time to interpret and respond to threats. AI may improve intelligence and early warning, but it can also introduce errors that appear authoritative. Automated systems may accelerate provocative military decisions precisely when caution is most necessary.
It is not hard to imagine an AI-assisted warning system misidentifying an incoming nuclear attack or a cyber intrusion into nuclear command infrastructure as preparation for a first strike. Autonomous weapons could escalate a conventional confrontation before political leaders fully understand what has happened.
AI could also enable sophisticated attacks on communications and critical infrastructure, increasing uncertainty about an adversary’s intentions. In short, the danger is not simply that AI makes weapons more powerful but that it undermines the predictability, communication and human judgment that nuclear deterrence depends on.
The urgency is growing. New START, the last treaty limiting American and Russian strategic nuclear forces, expired in February 2026 without a replacement. China’s expanding nuclear arsenal further complicates strategic stability. Meanwhile, AI capabilities evolve faster than traditional arms-control institutions.
A comprehensive trilateral agreement would be unrealistic. Washington and Moscow possess much larger nuclear arsenals than Beijing, which fears that transparency could expose vulnerabilities in its deterrent. Russia relies heavily on nuclear weapons for its great-power status and would be less inclined to diminish its arsenal.
That is why the three powers should begin with something more modest but potentially consequential: a standing trilateral dialogue on AI and nuclear strategic stability. Such a mechanism would bring together diplomatic, military, nuclear-security and technological officials, supported by scientific experts. It would build on existing discussions, including the American and Chinese commitment in 2024 to maintain human control over decisions to use nuclear weapons.
Its initial objectives should be practical: protecting nuclear command-and-control systems against cyber interference, maintaining reliable crisis communications, preserving human authorization over nuclear use and establishing procedures for reporting dangerous incidents involving autonomous systems. Over time, consultations might produce common definitions, confidence-building measures and limited reciprocal restraints.
Success would be far from guaranteed. Mutual suspicion is profound, verification is difficult and dialogue could become diplomatic theater. The United Kingdom, France, India and other nuclear powers would eventually need to participate in broader arrangements. The initial trilateral mechanism should not grant Washington, Beijing and Moscow exclusive authority over humanity’s security.
Nevertheless, even limited progress would matter because it could satisfy all three dimensions of international stability within a narrowly defined field. In power, it would recognize the countries’ distinctive nuclear and technological capabilities. In rules, it could establish reciprocal restraints against catastrophic miscalculation. In values, it could affirm a minimum principle transcending incompatible political systems: no government should gamble with humanity’s survival.
Passing the AI-nuclear test is necessary, but not sufficient for stability. If the three powers cannot cooperate against dangers threatening their own existence, their prospects for constructing a stable international order are exceedingly poor.
Of course, preventing technological catastrophe would not resolve the underlying territorial disputes that threaten global peace. But the three leaders cannot build a more legitimate international order merely by dividing influence among themselves. They must take an indispensable first step in agreeing that humanity’s most dangerous technologies must remain subject to human judgment and common rules.

Mr.BadGuy
Publisher
Mr.BadGuy
Publisher
These soldiers are from a country whose president is a Jew Zelensky. Orthodox churches are also being destroyed in this country. Synagogues are not touched. Surprisingly, it turns out that Jews get along well with the Bandera Nazis.