At the invitation of Chinese President Xi Jinping, U.S. President Donald Trump paid a state visit to China on May 13-15, his first trip to the country in nine years. The two heads of state held in-depth discussions on bilateral, regional and global issues. AI governance and strategic stability stood out as areas where the international community expected to see meaningful China-U.S. consensus.

Rising risks,narrowing window

AI is moving rapidly into high-risk domains, from autonomous weapons and unmanned systems to nuclear decision-making, turning once-theoretical threats into real-world challenges.

Some in the United States have called for developing nuclear-capable autonomous systems and deploying them in the Asia-Pacific region. AI-enabled unmanned platforms, from underwater vehicles to drones, can combine stealth, speed and operational unpredictability. A single miscalculation, malfunction or cyberattack could trigger escalation, with potentially irreversible nuclear consequences. The greater danger is that once such a system makes an escalatory move, human operators may have too little time to intervene, or may be cut out of the decision-making loop altogether.

Against this backdrop, Trump's visit offered a crucial opportunity. China and the U.S. should urgently work toward prohibiting the integration of nuclear weapons with autonomous systems and establishing dedicated risk-management channels for unmanned platforms. Responsible governance of AI's military and security risks is no longer optional, it is a shared imperative.

Getting AI governance right

AI's impact extends far beyond security. It is being woven into economic and social life at unprecedented speed, transforming how societies produce, deliver services and govern. From advanced manufacturing and smart healthcare to education, urban management, scientific research and public services, AI has become a major driver of today's technological and industrial transformation, reshaping global competition, industrial structures, as well as governance systems. No country or region remains untouched by this transformation.

AI transcends borders, platforms and sectors. Consequently, challenges like cross-border data flows, model training, computing infrastructure, supply chains, standard-setting and risk spillovers cannot be addressed by any single country or region alone. Excessive data collection, algorithmic echo chambers, automated decisions that entrench discrimination, deepfakes that erode public trust and AI tools used for cyberattacks, unfair competition, or even military conflict, all are symptoms of a technology that has outgrown the frameworks designed to govern it. As AI becomes more deeply intertwined with data, computing power, platforms and weapons systems, it is now a major concern for economic security, social stability, ethical norms and international peace.

On this front, Chinese and U.S. lawmakers have identified a broadly similar set of concerns: protecting personal data, regulating data processing, ensuring algorithmic fairness, guarding against discriminatory automated decision-making, securing user rights, holding platforms accountable, safeguarding minors and managing high-risk applications. The two countries' legal frameworks differ in form, but their underlying objectives are broadly aligned: to promote the safe and beneficial development of AI and ensure that it advances the public good rather than undermining human dignity, rights or safety. That is what getting AI governance right looks like.

What makes AI governance "good?" It begins with a simple principle: Innovation must serve human wellbeing. AI can truly serve society rather than dominate it only when it is developed and deployed on human terms, with human judgment, oversight and accountability preserved at every pivotal decision point. Both countries should direct AI development and deployment toward areas that improve everyday life: easing clinicians' diagnostic workloads, improving teaching and learning, helping workers adapt to industrial transformations, strengthening public services and expanding support for vulnerable populations.

What should governance aim for? Its purpose is to create a clear, stable and credible regulatory framework for innovation. Effective governance means that data is collected lawfully, that sensitive information is properly safeguarded and that algorithmic processes are fair, transparent and explainable. It means safe and reliable model outputs, corporate accountability, effective redress for users and protection of the public interest against technological abuse. The more powerful AI becomes, the more robust its governance must be, across regulation, accountability and ethics alike.

Shared risks to shared rules

Good AI governance must also be built on a fundamental principle: Safety and innovation must advance together. Without adequate safeguards, innovation risks drifting off course. Both China and the U.S. are home to world-leading foundation models, computing resources and AI ecosystems, and both face a similar set of complex challenges: model hallucinations, data breaches, algorithmic bias, malicious use, among others. For AI systems with substantial social impact, the two countries could pursue joint approaches to risk classification, safety assessment, red-teaming, ethical review, content labeling, deepfake regulation and the protection of minors. Especially regarding high-risk applications, both sides should work toward establishing clearer safety thresholds and rules of accountability, ensuring that AI is not used to endanger public safety, violate fundamental rights or undermine international peace.

Development and security are not at odds. Sound regulation strengthens public trust and supports long-term industry growth. Meanwhile, longstanding challenges such as the digital divide demand equal attention, and must be addressed with equity and inclusiveness in mind.

As two major players in AI, China and the U.S. should adopt a more open approach to capacity building, cooperating on talent development, infrastructure, open-source ecosystems, public goods and the deployment of proven applications, while respecting each other's sovereignty and development path.

The two countries should deepen coordination in multilateral forums and pursue practical cooperation through bilateral channels, establishing regular dialogues on AI safety, mutual recognition of standards, ethical and regulatory frameworks, and risk assessment. Where the two countries compete, they should do so responsibly; where shared interests are greatest, cooperation should run deepest.

Both sides should seek common ground while shelving differences, and safeguard their own interests without losing sight of the global good. The goal is not merely to manage AI, but to jointly shape the governance framework for the age of AI.

The author is a professor of law at the Beijing Institute of Technology and executive director of the institute's Research Center for Science and Technology and Human Rights

Copyedited by Elsbeth van Paridon

Comments to dingying@cicgamericas.com