As President Trump prepares to travel to Beijing for meetings with Xi Jinping on May 14-15, a new commentary from the Brookings Institution argues that both countries have the opportunity and strategic incentive to establish practical cooperation frameworks in artificial intelligence, even as they remain locked in fierce technological competition. The paper, authored by Kyle Chan and Ryan Hass, outlines three concrete proposals that could form the basis of the first structured bilateral AI risk dialogue between the world’s two leading powers.
The timing of the commnetary is deliberate. Treasury Secretary Scott Bessent’s team is actively preparing for the Beijing summit, and Treasury officials have signaled that Trump and Xi plan to discuss AI specifically, with a focus on “security and threats from nonstate actors.” Even a nonbinding AI safety declaration, the authors argue, would mark a historic milestone, the first structured bilateral framework on AI risk between the US and China.
The Case for Cooperation Amid Competition
The central argument of the Brookings commentaryis that AI cooperation and AI competition are not mutually exclusive. The United States and China can continue to compete vigorously in developing frontier AI models, building AI-powered military capabilities, and dominating global AI markets, while simultaneously taking practical steps to reduce the shared risks posed by unchecked AI development to both countries.
The authors identify the most acute shared risk as the prospect of an AI arms race in which both countries feel pressure to integrate AI into military operations as quickly as possible, potentially at the expense of safety precautions and human oversight. This dynamic, they argue, could lead to catastrophic miscalculation, not through deliberate aggression, but through the failure of AI systems operating faster than human decision-makers can respond.
A second shared risk is the misuse of AI by nonstate actors. Terrorist organizations, criminal networks, and rogue states could exploit powerful AI systems to develop biological or chemical weapons, conduct large-scale cyberattacks, or execute sophisticated disinformation campaigns. Neither the US nor China has an interest in enabling these outcomes, and both have strong incentives to cooperate in preventing them.
Three Concrete Proposals
The Brookings paper outlines three specific mechanisms for US-China AI cooperation, each designed to be politically feasible and practically implementable without requiring either side to compromise its core competitive interests.
The first proposal is the development of nonbinding AI guidelines, common safety guardrails for frontier models that would include shared definitions of high-risk use cases, baseline restrictions on dangerous model behavior (such as providing detailed instructions for creating weapons of mass destruction), and commitments to prevent “safety arbitrage,” where AI developers in one country exploit the absence of safety standards in another to gain competitive advantages.
The second proposal involves limited information sharing on AI-related security threats. This would not require either country to reveal proprietary model architectures or training data but would involve sharing examples of prompts, outputs, and behavioral patterns associated with malicious use. The goal is to build a shared understanding of how AI systems are being exploited by bad actors, enabling both countries to develop more effective countermeasures.
The third proposal is the establishment of an AI emergency hotline, along with formal and informal communication channels that could be activated in the event of an AI-related incident. The authors draw an analogy to the nuclear hotlines established between the US and Soviet Union during the Cold War, which proved invaluable in preventing miscalculation during crises. An equivalent mechanism for AI would provide a critical safety valve in a world where AI systems are increasingly integrated into critical infrastructure and military command-and-control systems.
The Motivations on Both Sides
The Brookings commentary is notable for its candid assessment of the strategic motivations driving each country’s interest in AI cooperation. For China, the primary motivation is to position Xi Jinping as a global statesman willing to engage constructively with the United States on shared concerns. By agreeing to even a modest AI safety framework, China can signal to the rest of the world that it is a responsible actor in the AI domain, potentially strengthening its hand in multilateral AI governance discussions.
For the United States, the motivation is to reassert American leadership in global AI governance before China unilaterally defines the terms. As we reported in our analysis of how China is quietly building a global AI standard, Beijing has been actively shaping international AI norms through multilateral forums, often in the absence of meaningful US engagement. A bilateral framework with China would give Washington a direct role in shaping the rules of the road.
Xi may also seek what the authors describe as “a floor” on US containment efforts, a commitment that AI cooperation will not be used as a pretext for further export controls or sanctions targeting Chinese AI companies. This is a sensitive issue given the ongoing restrictions on Huawei’s Ascend chips and the broader US campaign to limit China’s access to advanced semiconductor technology.
The Obstacles Ahead
The Brookings commentary acknowledges significant obstacles to any US-China AI cooperation framework. The bilateral relationship remains defined more by the absence of active friction than by any affirmative shared agenda. The structural tensions are real: the United States has imposed sweeping export controls on advanced AI chips targeting Chinese companies, while China has responded with its own restrictions on critical minerals and rare earth exports. These competing pressures make the diplomatic environment for AI cooperation genuinely difficult.
There is also a fundamental asymmetry in how the two countries approach AI governance. The United States has historically favored light-touch regulation and market-driven innovation, while China has developed a comprehensive state-directed approach to AI development and deployment. Bridging these different regulatory philosophies will require sustained diplomatic effort that goes well beyond a single summit meeting.
Nevertheless, the Brookings commentary hints at the alternative: allowing the AI relationship to drift toward pure competition without any guardrails is more dangerous. The question is not whether the US and China can afford to cooperate on AI, but whether they can afford not to.
As Trump and Xi prepare to meet in Beijing, the answer to that question may shape the trajectory of the most consequential technology of the 21st century.
(Related: Trump and Xi Bring AI Risks to Beijing as Summit Agenda Takes Shape)