AI can speed up development of Global South

Editor's note: Artificial intelligence can speed up the growth of developing countries. But these countries should take proactive steps to fully capitalize on the opportunities. People's Daily Overseas Edition spoke to Xu Xiujun, deputy head of the National Academy of Chinese Modernization at the Chinese Academy of Social Sciences, and Jiang Tianjiao, associate professor at the Development Institute of Fudan University, about the challenges the Global South faces in leveraging AI and how international cooperation can help overcome these hurdles. Below are excerpts of the interview. The views don't necessarily represent those of China Daily.

AI technologies can diffuse through open source frameworks, pretrained models and cloud-based application programming interface calls. This allows developing countries to avoid costly hardware and gain direct access to advanced algorithmic capabilities.

Developing countries don't have to develop state-of-the-art AI models from scratch. Instead, they can adopt existing tools and tailor them to local conditions. AI applications in agriculture, healthcare, education, public governance, disaster response and other fields can enhance public services. This way, AI can ease the shortage of professionals and public services in some developing countries.

The competition over AI involves a comprehensive ecosystem that integrates data, algorithms and computing power. Currently, a few countries and enterprises dominate advanced computing resources, which raises barriers for developing countries. At the same time, many Global South countries face challenges such as insufficient power and network infrastructure, limited access to computing power, lack of localized data and talent, as well as underdeveloped industrial ecosystems for research and development and the application of AI. This threatens to turn the "access divide" of the internet era into an "intelligence divide".

AI has also given rise to a new "representational divide". Large-model training relies heavily on data, yet languages, cultures and knowledge of many developing countries are under-represented in the training datasets of mainstream models. This could lower the suitability of models and lead to algorithmic bias, undermining the performance of such AI models in local contexts.

The AI era requires not only basic digital literacy but also skills in human-machine collaboration, placing greater demands on education, vocational training and talent development. A country's ability to build a talent base compatible with intelligent transformation will determine whether it can fully unlock the dividends of AI.

Without an adequate understanding of AI application and regulation, and without policies and rules for AI development, governments risk long-term dependence on foreign technology suppliers and large platform enterprises. This dependence could hinder the development and growth of domestic industrial ecosystems and innovation capacity.

Future international cooperation should focus on systematically enhancing the capabilities of Global South countries. Priority should be given to developing inclusive digital infrastructure, including the power supply, networks and computing capacity, to support Global South countries in building more local, green and resilient intelligent computing centers. Efforts should also be made to establish an open-source and inclusive technology ecosystem. This includes encouraging the development and application of open-source AI models, frameworks and tools, and supporting Global South countries in cultivating local talent, native language models and AI applications tailored to their local needs.

Furthermore, countries should cooperate to establish a fair and reasonable global AI governance system. Platforms such as the newly established World Artificial Intelligence Cooperation Organization should be leveraged to ensure Global South countries participate on an equal footing in formulating international AI rules, technical standards and ethical norms. Priority should be given to cooperation in areas related to people's well-being, including agriculture, healthcare, education, and disaster early-warning, to ensure that AI dividends can truly benefit broader populations.