As a fast-evolving and disruptive technology, AI is showing ever more clearly its nature as a double-edged sword. It offers vast opportunities for social and economic progress, but also carries significant uncertainty and complex risks. Addressing these uncertainties and risks, and ensuring AI develops in an inclusive and responsible way, requires urgent global dialogue and cooperation.

Multi-faceted challenges

Global AI governance faces multiple challenges. Chief among them are the risks inherent in the technology itself. A prominent example is the “black box” problem: For a given input, an AI system generates results through complex algorithms and models, yet users are often unable to explain how those results are derived or to grasp the details of the underlying mechanisms. Put simply, the inner workings of AI remain opaque. Another concern is “hallucination,” where AI produces false or misleading information. AI systems may invent fictional storylines, speculate about user preferences, or even fabricate references and sources, all with convincing detail. The misuse of AI is also a significant risk—for instance, the deployment of AI-enabled weapons in regional conflicts can trigger humanitarian crises, while criminals are exploiting deepfake technology to commit fraud, spread disinformation, and incite violence.

Another major challenge lies in the imbalance of AI capabilities, which has deepened the global digital divide. A handful of developed countries and leading corporations monopolize the core resources and technologies essential to AI and, driven by geopolitical and economic motives, often block technology transfer or impose discriminatory standards that deny developing countries fair opportunities. At the same time, developing countries suffer from inadequate infrastructure and limited technical capacity, leaving them heavily reliant on external technologies and trapped in a structural dilemma of starting behind and staying behind. The persistence and widening of this intelligence gap not only consigns the Global South to the bottom of global value chains but also erodes their voice in shaping AI agendas and standards. Worse still, it has fostered new forms of governance inequality, marked by algorithmic dependency and digital colonialism.

The third challenge is that global mechanisms for AI governance remain underdeveloped. While international organizations and platforms such as the Organization for Economic Co-operation and Development, UNESCO, the Group of Seven (G7) and the Group of 20 (G20) have launched initiatives to build governance frameworks, these efforts remain fragmented—differing in implementation pathways, priorities, and participation mechanisms. Policies and standards vary sharply across jurisdictions; and no unified, binding framework has yet emerged. For instance, the United States is moving toward deregulation, whereas the European Union has established a rules-based system through hard law, leaving standards difficult to align. This lack of coherence hampers mutual recognition and the interoperability of governance rules, creating regulatory disorder and barriers to collective action.

Providing public good

Confronted with the complexities of global AI governance, China is taking an active lead, working to ensure that AI develops in an inclusive and beneficial direction.

China adheres to the core principle of “people-centered, AI for good” in advancing AI development and governance. In October 2023, President Xi Jinping proposed the Global AI Governance Initiative, emphasizing the need to balance development with security, foster dialogue and cooperation to build consensus; and establish an open, fair, and effective governance framework that enables AI to serve the wellbeing of humanity. In September 2024, China introduced the AI Capacity-Building Action Plan for Good and for All (ACBAPGA) at the United Nations High-Level Meeting on International Cooperation on Artificial Intelligence Capacity-Building, setting out five key principles: sovereign equality, orientation toward development, people-centeredness, inclusiveness and collaborative cooperation. In July 2025, China released the Global AI Governance Action Plan, calling on all parties to uphold the goals and principles of promoting AI for good and for the people, pursuing development, ensuring safety and controllability, and advancing fairness and inclusiveness, while taking concrete and effective actions to jointly promote global AI development and governance. This August, China’s State Council issued the Opinions on Deepening the Implementation of the AI Plus Initiative, explicitly defining AI as an international public good for the benefit of humanity and calling for the creation of an open ecosystem for AI capacity-building that is equal, trustworthy, diverse and mutually beneficial. This series of initiatives proposed by China has increased international consensus on advancing AI for the common good, contributing Chinese wisdom and solutions to global AI governance.

As a responsible major country committed to advancing AI development in a people-centered and beneficial manner, China supports the United Nations in playing the central role in global AI governance. It actively promotes UN principles in both bilateral and multilateral settings and works to translate UN initiatives into concrete outcomes. In June 2024, with joint sponsorship or support from China, the United States and 120 other countries, the 78th UN General Assembly adopted its first-ever resolution on AI, calling for it to benefit all countries and regions, bridge the digital divide, advance digital transformation, and promote peace. In July 2024, the General Assembly unanimously adopted a resolution put forward by China on strengthening international cooperation in AI capacity-building, which underscores that AI development should remain people-centered, oriented toward good, and dedicated to serving humanity. The resolution also reaffirms support for the UN’s central role in international cooperation to achieve inclusive, beneficial, and sustainable AI development. As a key promoter, China has fulfilled its responsibilities by introducing the ACBAPGA as a measure to enhance implementation of the UN General Assembly resolution on strengthening international cooperation in AI capacity-building. The plan is being advanced through North-South cooperation, South-South cooperation, and triangular cooperation. In addition, China co-hosted with the UN Secretariat the first AI Capacity-Building Workshop in Shanghai, which drew representatives from nearly 40 countries and achieved positive results.

Unlike the closed-source model dominated by U.S. giants such as OpenAI, China has adhered to the?principle of openness and open source in the field of AI. It has put forward the International AI Open-Source Cooperation Initiative, aimed at building an open-source ecosystem in partnership with other countries, with open source serving as the key driver. This reflects China’s commitment to promoting and leading an open, inclusive, and beneficial path of AI development. For example, DeepSeek-R1, with its open-source, low-cost and high-performance model, has provided a new paradigm and direction for global AI development. It has lowered the threshold for AI adoption, facilitated technology diffusion and advanced the democratization of innovation, thereby helping to bridge AI gaps between the Global North and South, as well as across regions. China has also actively shared technological achievements and best practices with its global partners, creating platforms for cooperation and organizing a wide range of exchanges to foster a multi-level and diversified ecosystem of collaboration. In healthcare, for instance, AI has been applied to intelligent medical imaging analysis, improving patient diagnosis and treatment experiences. In the energy sector, the Bright Power large model has been applied in Brazil to enhance the quality and efficiency of grid inspections and to strengthen power grid security.

AI cooperation in the Global South

China has consistently paid close attention to the development needs of Global South countries in AI and has taken measures to advance robust cooperation with them in this domain.

China has worked through platforms such as BRICS (a platform for cooperation between emerging-markets that initially comprised Brazil, Russia, India, China and South Africa) and the Shanghai Cooperation Organization (SCO) to establish and improve cooperation frameworks on AI. Within the BRICS framework, China promoted the release of a joint initiative to strengthen cooperation on AI application and development, thereby enhancing the voices and cooperation capacities of member states. The Rio de Janeiro Declaration stressed that global AI governance should focus on reducing risks, responding to the needs of all countries—including those of the Global South—building a UN-centered governance system, and promoting the responsible development and application of technologies in support of sustainable development. Within the SCO, China has promoted the adoption of the Tianjin Declaration, which supports greater cooperation among science parks and innovation clusters, and advanced the formulation of the SCO Outline on Future Technologies. It has also leveraged the potential of the Alem.AI International AI Center to establish mechanisms for AI cooperation. Moreover, China issued a plan for building the China-SCO AI Application Cooperation Center, inviting other SCO member states to jointly build a cooperation center for AI applications, to promote inclusive and widespread use of AI?technologies.

China has deepened practical cooperation through initiatives such as the China-BRICS AI Center and capacity-building workshops. The center was established to pool resources, promote research and development, facilitate exchanges, and nurture talent, providing vital support for Global South countries. In 2024, China and Zambia jointly launched the Group of Friends on International Cooperation in AI Capacity-Building. Earlier this year, China and Brazil co-hosted the BRICS Future Network Innovation Forum, which set out the responsibilities of research groups on AI and related areas. China has also been an active participant in the UN capacity-building network, helping developing countries cultivate expertise in AI science and applications through workshops that range from foundational theory to hands-on practice.

China has advanced practical cooperation through joint efforts in computing power, algorithm sharing and governance consultation. On computing power, China has prioritized development and inclusiveness by supporting Global South countries in strengthening infrastructure and building application scenarios. For example, in May this year, Malaysia launched its National AI Infrastructure Strategy, creating a full-stack AI ecosystem based on Huawei’s Ascend chips and the DeepSeek model. On algorithms, Chinese research institutions and enterprises have shared proven solutions to help Global South countries tackle real-world challenges. For instance, the AI visual hub system of Chinese company Vietadata integrates more than 50 algorithms with digital twin technologies to provide real-time monitoring and automated management of the Djibouti Free Trade Zone. On governance, China has called for Global South countries to take part in rule-making, respecting national contexts and development needs while working together to build a fair and effective global governance system. The Shanghai Declaration on Global AI Governance advocates establishing a global governance mechanism, reinforcing the UN’s central role, and enhancing the representation and voice of developing countries.

In global AI governance, China has set an example by steering technology toward the good, strengthening cooperation with the Global South, and contributing Chinese knowhow and solutions to building a community with a shared future for humanity. Looking ahead, China will continue to work with the international community to uphold openness and sharing, build platforms for cooperation, refine policy frameworks, promote technological collaboration and knowledge transfer, and lower barriers to innovation—jointly advancing AI development and ensuring it serves the common good.

The author is a research fellow with the Center for Global AI Innovative Governance and deputy director of the Center for BRICS Studies at the Development Institute of Fudan University in Shanghai.