New Global Order for AI Safety Governance: Breaking Unilateral Confrontation with Solidarity

In the wave of global technological competition, artificial intelligence has become a key force reshaping human society. However, along with the technological dividends come unprecedented safety risks and governance challenges. An advisor to the Legislative Research Society of the China Law Society, an expert advisor to the Peking University Center for Law and AI Research, and a Chinese expert at the Digital Economy Sub-Forum of the Beijing-Tokyo Forum noted that on July 17, President Xi Jinping delivered an important speech at the opening ceremony of the World AI Conference and High-Level Meeting on Global AI Governance, precisely addressing the era's questions on AI governance. This not only sets the tone for China's AI development path but also presents a profoundly inspiring Chinese approach to global governance.

1. The Era's Challenge for AI Governance: Balancing Development and Security

The widespread adoption and rapid iteration of AI technology have made it a highly pervasive frontier technology. China has consistently adhered to the principle of balancing development and security, proactively grasping AI development trends and laws, and continuously improving relevant laws, regulations, and ethical guidelines to ensure AI is safe, reliable, and controllable. However, the innovation dividends and associated risks of AI also affect the entire world. Ensuring AI safety, controllability, and strengthening risk governance are important issues facing all countries.

2. Creating a New Paradigm for Global Governance with the Spirit of Global Solidarity

First, the international community must embrace the spirit of global solidarity and establish a new paradigm for global governance through international coordination on AI. Notably, on July 16, the day before the opening of the 2026 World AI Conference (WAIC 2026), the World AI Cooperation Organization, with 29 founding member states, was established in Shanghai. The organization adheres to the purposes and principles of the UN Charter, respects national sovereignty and diversity of civilizations, upholds equal treatment, practices multilateralism, follows the principles of extensive consultation, joint contribution, and shared benefits, and is open to all countries.

It is foreseeable that the World AI Cooperation Organization, as an open international coordination platform for AI governance, will create a new paradigm for global AI safety and governance, pooling wisdom from all sides to jointly address the challenges of the AI era.

3. Resolutely Opposing the Politicization of Safety Issues and Technology Blockades

Second, the practice of certain countries that generalize, politicize, or even weaponize AI safety issues must be resolutely opposed. Currently, some countries, under the pretext of national security, abuse regulatory measures, impose technology blockades and model supply interruptions on others, forcing the global AI industry chain toward decoupling and fragmentation. Such artificial blockades and splits not only magnify AI safety risks but also erode the trust foundation of global governance, forcing other countries to develop alternative technology stacks, ultimately leading to incompatible technical standards, data formats, and security protocols between countries and regions. The more AI technology develops, the more we must break down barriers to safety governance regulation and resolutely oppose artificial decoupling and fragmentation.

4. Promoting International Standard Setting and Shared Infrastructure Construction

Third, we must actively advance the formulation and revision of international standards for AI safety, industry, and ethics. Specifically, we should balance technological progress, risk prevention, and social ethics, leverage the platform of the World AI Cooperation Organization, and gradually form a framework of standards and norms that is broadly consensual, scientific, transparent, and inclusive, promoting mutual recognition of standards among countries and reducing the cost and threshold of international AI cooperation. Since 2023, China has played an important leading and driving role in advancing AI safety governance standards. The 15 core consensus and action initiatives announced at this conference have been recognized and supported by many countries.

Fourth, build globally shared public infrastructure for AI risk governance. For example, establish cross-border AI risk early warning sharing platforms for real-time synchronization of deepfake samples, malicious models, algorithm vulnerabilities, and other intelligence among countries. At the same time, open trustworthy AI assessment tools and open-source risk control models to developing countries to address governance capability gaps. In addition, promote the AI governance capacity universal plan, providing safety standard training and computing power support to Global South countries to bridge the AI governance capacity gap.

5. Strengthening Safety and Governance Cooperation in East Asia

Fifth, strengthen AI safety and governance cooperation in East Asia. China, Japan, and South Korea are all major digital economies in East Asia with large AI industry scales and regulatory approaches that share consensus and complementarity, making it necessary and possible to form a pioneer region for international cooperation on AI safety and governance. In the near term, we can jointly build a mechanism for mutual recognition of AI safety standards, jointly participate in the drafting of international AI standards, and unify technical indicators for deepfake detection, algorithm bias assessment, and large model safety evaluation. At the same time, create a cross-border AI risk early warning and joint prevention platform to conduct joint analysis of cross-border telecom fraud, AI attacks on critical infrastructure, etc.

On July 16, the new generation model Kimi K3 was launched, with 2.8 trillion parameters, making it the world's largest open-source model by parameter count, marking a new step in China's AI model development. Such technological breakthroughs need the backing of international cooperation.

6. Conclusion: Jointly Building a New Pattern of Good AI Governance

Artificial intelligence has no borders, and safety risks also have no borders. Unilateral confrontation, fragmented rules, and generalized security thinking will only exacerbate common risks for all humanity. Looking ahead, we must firmly grasp the main line of balancing safety, controllability, and development. Domestically, establish a comprehensive risk prevention and control system across the entire chain to build a trustworthy local AI ecosystem; externally, break the shackles of unilateral technology blockades and promote the implementation of globally unified trustworthy AI standards. Only by coordinating domestic and international efforts, safety controllability, and risk governance can we jointly build a new global pattern of AI good governance that is people-centered, intelligent for good, fair and inclusive, and collaboratively governed.