As artificial intelligence rapidly advances while exposing widening global capability gaps, China on Tuesday called for strengthened international cooperation on AI capacity building at the United Nations.

Speaking at a meeting on "AI Empowering Engineering Capacity Building" at UN headquarters in New York, China's vice-minister of science and technology, Chen Jiachang, said China is promoting global AI governance within the UN framework and advancing international cooperation through initiatives such as an AI capacity-building resolution adopted by consensus at the UN General Assembly.

He said China is pursuing an open-source, innovation-driven approach to artificial intelligence and aims to foster a more inclusive global innovation ecosystem. Chen noted that China has become the leading source of AI patents worldwide, contributing about 60 percent of the total. He added that industry-led open-source AI communities in China have developed more than 170,000 multimodal models and now support over 20 million developers globally.

The meeting, convened under the Group of Friends for International Cooperation on AI Capacity Building co-chaired by China and Zambia, focused on practical cooperation to support developing countries in strengthening AI capabilities.

China's permanent representative to the United Nations, Fu Cong, who chaired the meeting, said global AI governance must keep pace with technological change and ensure that the benefits of the technology are broadly shared.

"It is essential that all countries, especially those from the Global South, can effectively participate in global governance, and that AI does not become the preserve of rich countries or wealthy elites," he said.

Over the past two years, the group has held four open meetings and three workshops in China, bringing together nearly 200 representatives from member states. New initiatives are planned in cooperation with the UN Department of Economic and Social Affairs, including regional workshops in Africa and the Asia-Pacific, Fu said.

According to a report published last month by the Stanford Institute for Human-Centered Artificial Intelligence, leading AI systems developed in China and the United States have "effectively closed the performance gap," even as access to computing power, data and talent remains highly concentrated and unevenly distributed.

Participants of Tuesday's event from developing countries also highlighted concerns that unequal access to infrastructure and expertise could deepen existing divides.

Zambia's permanent representative to the United Nations, Chola Milambo, said gaps in computing power, data and financing risk widening disparities between developed and developing economies.

"For Zambia and for many of our partners across Africa and the wider Global South, the question is no longer whether to engage with AI, but how to ensure that our engagement is inclusive, equitable and firmly anchored in our own development priorities," Milambo said.

"Capacity building is not a transfer, it is a partnership," he said, calling for development approaches aligned with national priorities.

Participants said strengthening cooperation on open technologies, talent development and infrastructure will be key to ensuring that AI contributes to inclusive and sustainable development, particularly in the Global South.