The Bishkek Declaration adopted at the Shanghai Cooperation Organization (SCO) summit in the capital of Kyrgyzstan last month calls for more equitable global governance and safe, responsible cooperation in artificial intelligence. China also advocated closer collaboration in AI and the digital economy.
The significance of these proposals lies not only in the technology itself but also in the question they raise: As AI reshapes economies and societies, will developing countries be able to participate meaningfully, or will they once again be left on the margins of a transformative tech revolution?
The World Bank’s World Development Report 2026 suggests that AI could compress the economic catch-up process for developing countries from a century to a decade, provided they act strategically. Unlike earlier tech revolutions, which rewarded countries capable of developing and controlling large segments of the value chain, AI offers a more accessible point of entry. Through open-source models, cloud-based services, and application programming interfaces (APIs), countries can draw on advanced capabilities without having to build frontier models from scratch. Adoption and adaptation can serve as effective entry points to development.
Developing economies are already playing an increasingly important role across the AI value chain, from data annotation and localization to application development, sector-specific fine-tuning, and deployment in agriculture, healthcare, logistics, and public administration. However, the most advanced layers of the value chain, including frontier-model development, chip design, and hyperscale computing, remain concentrated in a small number of leading economies. The distribution of private capital reflects the same asymmetry. High-income economies possess the vast majority of global data-center capacity, while low-income economies account for only a negligible share.
For developing economies, this gap has implications far beyond the tech sector. It affects whether governments can deploy reliable and secure AI-enabled public services at scale, from digital administration and education to healthcare, agriculture, disaster response, and financial inclusion. It also influences whether countries can retain control over sensitive public data, develop tools in local languages, and tailor AI systems to their own social and economic contexts. This creates an uneven AI landscape. Developing economies can benefit from advanced systems without building them from scratch. Yet if powerful models, computing infrastructure, and technological capabilities remain concentrated in a handful of countries and companies, these economies risk becoming dependent users of technologies over which they have little influence. The challenge is becoming more acute as AI is increasingly drawn into geopolitical competition. Recent media reports, citing a draft letter from the US State Department, suggest that countries participating in US-led AI initiatives may be encouraged to make clearer choices regarding their technology partnerships. If access to advanced models, chips, cloud infrastructure, and technical cooperation becomes shaped primarily by competing geopolitical blocs, developing countries may face difficult choices and reduced flexibility in pursuing technology strategies aligned with their own development priorities.
This asymmetry places a responsibility on economies and firms at the tech frontier. Ensuring that the benefits of AI are broadly shared is not merely an act of charity. It is essential to building a stable and legitimate global tech order in which the majority of the world’s population has a genuine stake. An inclusive AI future will require progress in three key areas.
First, equitable access to computing resources and APIs is essential if developing economies are to benefit meaningfully from AI. Advanced models, application interfaces, and computing resources should be made broadly available to researchers, businesses, and public institutions on nondiscriminatory terms. This need not compromise commercial viability. A tiered approach, in which basic capabilities are provided free of charge or at concessional rates while premium functions remain commercially priced, offers one practical solution. Such a model would expand access to development-oriented applications, including agricultural advisory services, health triage, classroom support, and productivity tools for small and medium-sized enterprises, while preserving the revenue necessary to sustain frontier research. “Freemium” models already underpin many consumer AI products. The real question is whether leading economies and AI companies are willing to deploy them deliberately as instruments of equitable access. Otherwise, geopolitical and commercial barriers may reproduce the exclusions associated with earlier technological waves.
Second, developing economies require sustained investment in AI talent, research infrastructure, and public-sector readiness through both multilateral and bilateral channels. China’s pledge to provide 5,000 AI training and seminar opportunities to developing countries over the next five years, for example, illustrates how bilateral initiatives can foster talent development and strengthen AI capabilities across the Global South.
Third, human-centered ethical frameworks must evolve alongside technological advancement. Like every major technological shift, AI will create both winners and losers. Robust governance frameworks are therefore necessary to prevent excessive concentration of AI resources, safeguard individual rights, mitigate systemic risks, and promote a human-centered approach to tech development. The newly established World Artificial Intelligence Cooperation Organization, comprising 29 member states, represents an effort to build a cooperative model for AI capacity-building and governance that enables countries in the Global South to benefit more equitably.
International cooperation, however, cannot substitute for the work that must take place within individual countries’ education systems. The capacity to understand, evaluate, and apply AI is rapidly becoming a foundational competence of the modern workforce, and education systems are being asked to adapt in real time. At every level, the goal is to cultivate the judgment necessary to use AI as a tool rather than as a substitute for independent thinking.
Up until now, higher education has led the way. Universities around the world are integrating AI into teaching, learning, and research at varying scales, ranging from dedicated AI programs to the incorporation of AI tools across disciplines. Yet higher education is too late a stage at which to begin. By the time students enter university, many of the foundational habits of thinking with, and about, intelligent machines have already been formed. AI literacy should therefore begin as early as possible, including at the primary-school level. A small but growing number of education systems have already adopted this approach. In Hong Kong, for example, the Education Bureau has established dedicated funding for primary and secondary schools to incorporate AI technology and content into both curricula and extracurricular activities, while also making continuing professional development in digital education mandatory for teachers.
The AI transition will ultimately be judged on two fronts: whether leading economies keep frontier capabilities accessible, and whether education systems equip the next generation to work with AI rather than around it.
How this transformation unfolds will be determined by the choices made today. Earlier technological revolutions were defined by those who developed the technology. However, AI’s legacy will depend on whether developing economies can harness its potential and help shape a more inclusive and prosperous future for all.
The author is president of Saint Francis University, Hong Kong.
The views do not necessarily reflect those of China Daily.
