China’s AI diplomacy is moving from declarations toward a more practical format. A six-day Artificial Intelligence Capacity Building Workshop opened in Beijing on September 7, bringing senior officials and core AI policymakers from 37 Global South countries to China for lectures, site visits, and discussions about governance and applications. The event was reported by Science and Technology Daily on September 8.
The workshop is hosted by China’s Ministry of Foreign Affairs and organized by Peking University, with support from Beijing municipal bodies. Participants include representatives from countries such as Kazakhstan, Russia, Algeria, and Brazil. The program is designed around a straightforward proposition: countries need more than access to AI products if they are to develop local capabilities and make choices about how the technology is governed.
That proposition is central to Beijing’s international AI message. China has repeatedly argued that the benefits of AI should not be concentrated in a small group of countries with the deepest pools of capital, computing power, and talent. The workshop does not create an international treaty or establish a new standards body. It is a capacity-building exercise, but one that gives China an opportunity to demonstrate the institutions, companies, research labs, and policy ideas it wants partners to see.
Training beyond a policy statement
The workshop’s structure indicates an attempt to make the offer tangible. Science and Technology Daily said the agenda includes nine specialized lectures, six site visits, and one roundtable. Its four main themes are AI safety governance, technology frontiers, application practice, and international cooperation.
The participants are expected to visit the Beijing Institute for General Artificial Intelligence, the Zhongguancun Exhibition Center, a humanoid-robot training center, and the science-fiction industry cluster at Shougang Park. Those visits matter because they show several layers of China’s AI ecosystem at once: fundamental research, commercial displays, robot training infrastructure, and creative industries that use AI tools.
The program also allows China to present its own development path as an example. It can show how municipal governments, universities, research institutes, and private companies interact, while making the case that a country’s AI strategy can be adapted to local economic conditions rather than copied directly from Silicon Valley or another major technology center.
The event follows earlier diplomatic efforts around a proposed World AI Cooperation Organization. That earlier effort concerned the construction of an international cooperation framework. The Beijing workshop is narrower and more operational: it is about giving officials exposure to policy tools, institutions, technical ideas, and use cases.
A focus on the global AI divide
Chinese organizers framed the program as part of the AI Capacity Building Inclusive Plan. In the opening discussion, officials and foreign guests considered what Chinese sources call the global “intelligence divide,” a reference to the risk that AI capabilities may be distributed unevenly between advanced economies and developing countries.
The concern has several dimensions. Developing countries may lack data centers, skilled personnel, high-quality local-language datasets, financing, and regulatory capacity. They may also depend heavily on systems developed elsewhere, with limited influence over how those systems are priced, governed, or adapted for local public services and businesses.
Training programs do not close those gaps by themselves. A week of lectures cannot substitute for sustained investment in electricity, connectivity, research institutions, and local entrepreneurs. Yet workshops can create networks among policymakers and expose them to potential partners. They can also influence how officials define the problem before they write national strategies or decide which foreign technology providers to work with.
China’s approach is notable because it combines technical exposure with a governance message. Participants are not only being shown laboratories and applications; they are also attending sessions on safety and global governance. That is a way of arguing that development and regulation should move together rather than treating safety as a concern reserved for wealthier countries.
What China gains from capacity building
The workshop serves China’s interests as well as the participants’. As Chinese firms expand cloud, model, hardware, and application businesses abroad, relationships with policymakers can influence where partnerships form and what kinds of local projects are considered feasible. Demonstrating China’s ecosystem may make it easier for foreign officials to evaluate Chinese technology on its own terms.
It also strengthens China’s argument that international AI governance should include countries beyond the United States, Europe, and a small group of other advanced economies. Building networks with Global South governments gives that argument a more practical foundation. The goal is not necessarily to persuade every participant to adopt Chinese products or policies. It is to make China a visible convenor in a conversation about who gets to build, use, and govern AI.
EastFrontier’s earlier examination of why U.S.-China AI cooperation may succeed where the trade war failed underscored how shared interests can coexist with deep competition. The Beijing workshop adds a third group to that picture: countries looking for practical ways to develop their own AI capacity while navigating competing technology ecosystems.
The measure of success will come later. It will be visible in follow-up exchanges, training partnerships, research links, procurement choices, and local AI programs, not simply in the workshop schedule. For now, the event shows that China is investing effort in the less glamorous work of building relationships and institutional familiarity around AI. In a global market where technology and governance are increasingly intertwined, that work can matter as much as a headline model release.