The Global South Chooses Its AI and Recomposes Digital Alliances
More than 40% of global ChatGPT traffic now originates from middle-income countries. Brazil, India, Indonesia, and Vietnam are leading this massive adoption of artificial intelligence tools that now exceeds that of developed nations. This silent shift redraws the geography of technological influence and transforms AI into a battleground between competing technological powers.
Washington and Beijing are exporting their algorithmic ecosystems with the same intensity they once deployed military bases. Each AI model adopted anchors a country within a sphere of technological influence from which it becomes difficult to escape. The Global South becomes the chessboard of this new digital cold war.
The Essentials
- More than 40% of ChatGPT users reside in emerging countries, with Brazil, India, Indonesia, and Vietnam leading
- 75% of developing countries depend exclusively on AI models designed in the United States or China according to the World Bank
- India is developing multilingual language models for several local languages with substantial investments
- China provides its AI tools free of charge to numerous African countries as part of its “Digital Silk Roads”
Emerging Countries Adopt AI Faster Than the West
The massive adoption of generative AI tools in emerging economies exceeds the most optimistic predictions. OpenAI records approximately 5.5 billion monthly visits to ChatGPT, a significant portion of which comes from middle and low-income countries. This proportion of 40% contrasts with the 23% of global population that these economies represent.
Brazil alone generates 147 million monthly requests, more than Germany and France combined. India follows with 134 million, surpassing the United Kingdom. This intensity of usage reveals considerable latent demand for cognitive automation technologies in economies where service sectors are experiencing rapid growth.
Structural factors explain this accelerated adoption. Emerging countries have young, connected populations, with 67% of 18-35 year-olds using digital tools daily according to World Bank data. Smartphone penetration reaches 85% in these economies, creating immediate infrastructure for accessing cloud-based AI services.
The absence of legacy infrastructure also plays a role. Unlike developed economies that must transform existing systems, emerging countries directly integrate AI into their new digital services. This logic of “technological leapfrogging” reproduces the phenomenon observed with mobile telephony, which allowed bypassing fixed networks.
Washington and Beijing Dispute Algorithmic Influence
This massive demand transforms AI into an instrument of technological diplomacy. The United States and China are exporting their models with radically different yet equally aggressive strategies. OpenAI, Google, and Microsoft favor domination through technical excellence and network effects. Their most advanced models capture the most sophisticated users, creating top-down dependence.
China bets on accessibility and industrial integration. Alibaba Cloud offers its AI tools free of charge in many African countries, accompanied by training and digital infrastructure. ByteDance deploys its recommendation algorithms through TikTok, creating complete ecosystems of integrated AI embedded in daily usage.
These strategies produce technological lock-in effects. A developer accustomed to OpenAI’s APIs structures their code according to American specifications. A company trained on Alibaba tools adapts its processes to Chinese standards. The initial choice of an AI ecosystem determines computing architecture for years to come.
The battle for talent amplifies these stakes. American universities train 23,000 AI PhD students from emerging countries each year, of which 40% remain in the United States after graduation according to the National Science Foundation. China counters by funding 150 AI research centers in 60 developing countries, training local engineers on its own technologies.
India and Brazil Attempt a Third Path
Facing this duopoly, some emerging powers are developing their own capabilities. India is investing substantial sums in language models covering multiple local languages and capable of processing cultural specificities that ChatGPT or Baidu neglect. The stated objective: reduce dependence on foreign giants for digital public services.
Brazil is launching “IA Brasil” with 800 million dollars, targeting agricultural and environmental applications where the country holds comparative advantages. This niche strategy aims to create specialized models not replaceable by generic American or Chinese solutions.
These national initiatives face economic realities. Developing a generalized AI model costs between 5 and 50 million dollars depending on its sophistication, not including continuous training costs. Technical talent is concentrated in a few global metropolitan areas, creating fierce competition to attract engineers capable of these developments.
The emerging pathway thus favors specialization and selective partnerships. India collaborates with Europe on certain open AI projects, while developing its own linguistic capabilities. Brazil integrates American models for financial services but develops its own algorithms for precision agriculture.
Digital Alliances Reshape Geopolitics
This technological fragmentation creates new blocs of influence. The American ecosystem federates English-speaking countries and economies most integrated into the Western financial system. Australia, Canada, the United Kingdom, and Singapore massively adopt OpenAI and Google tools, strengthening their strategic alignment with the United States.
The Chinese sphere structures itself around the “Digital Silk Roads.” Beijing provides infrastructure, training, and AI models to 68 partner countries, creating durable technical dependencies. Pakistan uses Chinese algorithms for its digital public services. Ethiopia deploys urban surveillance systems developed by Hikvision and powered by Chinese AI.
This geopolitics of AI reproduces the logic of traditional influence but accelerates it. Adopting an AI ecosystem determines technical standards, data protocols, evaluation criteria, and methods of algorithmic governance. A country that chooses Chinese tools implicitly integrates Chinese conceptions of privacy, surveillance, and social control.
Digital sovereignty stakes are sharpening. Europe is attempting to build its own path with models respecting its data protection standards, but struggles to compete with the attractiveness of American solutions or the accessibility of Chinese tools for emerging countries.
Africa, Laboratory of Algorithmic Diplomacy
Africa crystallizes these stakes. The continent has 1.4 billion inhabitants, sustained demographic growth, and immense needs for digital services. Chinese investments in African AI reach several billion dollars over five years, far exceeding American initiatives.
This asymmetry produces concrete effects. Kenya deploys digital public services developed by Chinese companies, integrating facial recognition algorithms and behavioral analysis. Nigeria uses Chinese language models adapted to local languages for its e-government platforms.
The United States counters through academic partnerships and training programs. USAID funds 47 AI innovation centers in 23 African countries, training local developers in American technologies. Microsoft and Google multiply agreements with African universities, offering free access to their cloud tools in exchange for research data and usage feedback.
This competition partially benefits African countries that negotiate advantageous conditions with both sides. Morocco obtains Chinese technology transfers for its “smart cities” while hosting Microsoft data centers for North Africa. South Africa develops partnerships with American universities while importing Chinese AI equipment for its mining sector.
Risks of Durable Fragmentation
This technological balkanization raises major challenges for global interoperability. AI models trained on different data, with divergent architectures and incompatible standards, create watertight digital silos. An algorithm optimized for Mandarin Chinese struggles to process technical English. A model trained on American data reproduces Western cultural biases.
The linguistic question aggravates these fragmentations. American models dominate in European languages but neglect the 2,000 African languages spoken by fewer than 10 million speakers each. Chinese tools excel in Mandarin and Cantonese but struggle with Arabic or Swahili. This linguistic distribution reinforces existing geopolitical spheres of influence.
Economic effects are already manifesting. Developers trained on OpenAI APIs struggle to migrate to Chinese alternatives and vice versa. Companies that standardize their processes on an AI ecosystem incur high transition costs to change suppliers. These technical frictions become de facto commercial barriers.
The issue of algorithmic governance also divides approaches. The United States privileges market regulation and industrial self-regulation, allowing companies to define their own ethical standards. China imposes strict state oversight with algorithms conforming to national political objectives. Europe develops a detailed regulatory approach but struggles to impose it beyond its borders.
Emerging Countries Face a Complex Technological Dilemma
For developing economies, this context creates a complex strategic dilemma. Adopting the most performant technologies guarantees immediate competitiveness but creates durable dependence. Developing national capacities preserves sovereignty but delays access to cutting-edge innovation.
Current choices will determine geopolitical balances in the decades to come. A country that massively integrates Chinese AI into its public services, education system, and critical infrastructure durably moors itself to the Chinese sphere of influence. Conversely, adoption of American ecosystems consolidates transatlantic and Western ties.
This battle for algorithmic influence silently redraws the map of alliances. Algorithms become digital ambassadors that propagate worldviews, social models, and technical standards. The Global South is not merely choosing technological tools — it is selecting its future geopolitical partners and its models of digital development.