Illustration: Chen Xia/GT
Recently, in a conference hall at the United Nations Headquarters in New York, a discussion on artificial intelligence (AI) development moved beyond algorithms and models and focused on the idea of a “shared future.” The Friends of International Cooperation Theme Conference on AI Capacity Building was co-chaired by representatives from China and Zambia, and attracted more than 120 representatives from more than 50 countries and international organizations. From basic computing power to algorithmic breakthroughs, engineering applications to deep space computing, AI technology is rapidly advancing, but also creating significant governance gaps. It is therefore urgent for the international community to ensure that the principle of AI for good and for all is effectively implemented.
As the world enters the AI era, security risks are no longer limited to the realms of science fiction. On the one hand, unprecedented opportunities exist. On the other hand, it creates unpredictable risks. Advanced AI models can solve complex problems and detect vulnerabilities at speeds and scales far beyond human capabilities, while also lowering the threshold for malicious use. Whether open source communities or closed source systems, they all face the difficult challenge of providing effective governance and preventing the global spread of risks. New York Times columnist Thomas L. Friedman once offered a vivid analogy. Two men in a cave with laptops and access to advanced AI models and Starlink terminals could potentially attack the critical infrastructure of any society. This is not an alarming story. Cybersecurity experts have warned that Frontier AI models have reduced the time needed to identify vulnerabilities in systems from weeks or even months to just minutes. Attackers only need to find one vulnerability, but defenders need to protect all vulnerabilities.
At a time when “governance has not kept up with the rapid development of AI” has become a common challenge across countries, the need to build a global AI governance framework based on broad consultation, collective contribution, and shared benefits has become even more urgent.
From deepfakes and cyberattacks to the use of AI for military applications in the recent Middle East conflict, the misuse of technology is alarming to humanity. No country is immune, and regulatory gaps in a single country can become a source of global risk. All stakeholders urgently need to take a broader view based on building a community with a common future for humanity, engage in dialogue on their respective regulatory and legislative frameworks and governance standards, and seek broad consensus while strengthening policy coordination on key issues such as AI safety and data flows.
On global AI governance issues, China’s position has been consistently proactive, responsible, open and consistent. Guided by the vision of a common future for humanity, China has actively implemented the Global Governance Initiative (GGI) and the Global AI Governance Initiative proposed by President Xi Jinping, and remains committed to contributing to global development and providing public goods. We have translated these visions into concrete actions by proposing and promoting initiatives one after another, including the “AI Capacity Building Action Plan for Good and All,” the “Global AI Governance Action Plan,” and the “AI+ International Cooperation Initiative.” In July 2024, the 78th United Nations General Assembly unanimously adopted a resolution on strengthening international cooperation on capacity building in artificial intelligence proposed by China, which was enthusiastically co-sponsored by more than 140 countries. This is a classic example of China contributing wisdom and solutions to global AI governance.
At the practical level, China is already in the lead. Through responsible top-level planning, the Chinese government is ensuring that open source large-scale models such as DeepSeek and MiniMax M2.5 are leveraged to power a wide range of industries, enhance public welfare, and promote international cooperation.
From universities in Brazil to financial institutions in South Africa, from helping build a relief bridge after Myanmar’s earthquake to supporting Cambodian fisheries transitioning from widespread fishing, China’s AI is contributing to its own high-quality developments while also benefiting the world through technological openness and co-building ecosystems.
It is the responsibility and obligation of all countries to make “AI for good” a reality. No single country, no matter how powerful, can dominate this vast field of AI or shoulder the global risks associated with this technology alone. Since the beginning of this year, more rational voices have emerged in the United States calling for Sino-American cooperation on AI. On May 6, The Wall Street Journal reported, citing sources familiar with the matter, that China and the United States are considering a broader, formal dialogue to consider AI governance. This is good news for the world. Indeed, the two countries have maintained collaboration on AI governance in recent years through both formal and Track II dialogues. However, to achieve further progress, it will be key for the United States to cooperate with China, abandon restrictions and repressive measures against China in the field of AI, and abandon its reluctant attempts to create an internationally exclusive bloc.
Zero-sum thinking and geopolitical conflicts are clearly the biggest obstacles to global AI governance. AI should not become a private property monopolized by certain powers, a tool of a closed club in certain countries, and certainly not a sword hanging over humanity.
Building a global AI governance system that is inclusive, fair, and sustainable is essential to the future and destiny of humanity. All parties should support true multilateralism and unite their efforts within the United Nations framework to achieve quality development. Sharing experiences and mutual learning is the key to overcoming governance dilemmas. We hope that all parties will be more open and collaborative than ever before, so that AI can become a bridge that connects the world and benefits society.
(Web editor: Zhong Wenxing, Liang Jun)
