Much of the Western discourse on Artificial Intelligence (AI) has lately focused on establishing safeguards and installing guardrails against powerful new AI systems, algorithmic bias, collusion of governments and tech oligarchs, and rising environmental costs related to AI ecosystems. The growing backlash in the West against the adverse effects of AI is labelled as ‘Botlash’ in the most recent commentary by Marietje Schaake. This commentary refers to various anti-AI movements that have gained tractions in the recent past including ‘QuitGPT’, ‘Resist and Unsubscribe’, and ‘Stealing Isn’t Innovation.’ While ‘botlash’ may be an apt description of how AI is now being perceived in the developed countries, the story for the Global South is completely opposite where AI is being viewed as some magical cure for poor governance, corruption, and weak economic development. Countries in the Global South are yet to undergo indigenous AI governance, development and deployment. Therefore, unlike the developed countries, the Global South is yet to experience a localised and large-scale adoption of AI or a ‘bot boom.’ However, the bid to adopt AI without first developing localised governance, digital literacy, and research ecosystem carries the risk of becoming passive consumers of foreign technologies.
AI is being pushed by political leaders and development agencies as the ultimate means of reigniting stagnant economic and development growth across multiple regions of Global South including Africa, South Asia and Latin America. Governments in these regions are presenting AI as a tool to fix bad governance, make healthcare and education more accessible, reduce corruption, and manage climate related disasters. For instance, in 2025 Ethiopia launched its Digital Ethiopia 2030 (DE2030) strategy, which explicitly calls for the integration of AI in education, healthcare, tax services, and justice. Similarly, Pakistan’s National AI policy 2025 frames AI as a transformative tool that would be employed in the sectors of healthcare, education, governance, agriculture and industry. Many countries in Latin America such as Chile, Argentina, and Colombia have also adopted national AI strategies that promote AI role in modernising public administration and fostering economic growth. However, none of these countries appear in the latest Stanford Global Vibrancy Ranking which means that they critically lack the pre-requisites for localised AI development and deployment including Digital Public Infrastructure (DPI), talent, research and development (R&D), data governance, and digital literacy.
AI development in countries of the Global South remains slow due to various structural asymmetries mainly in four domains including infrastructure, data, governance and literacy. Infrastructure asymmetry relates to the lack of substantial local AI R&D, advance computational power (data centres, GPUs), uninterrupted high-speed internet, surplus energy, and development of indigenous large language models (LLMs) in the Global South. In addition, overall public investment in AI related research also remains low in the Global South. For example, Pakistan’s R&D expenditure remains at 0.16 percent of the country’s total GDP. As a result, countries in the Global South not only risk falling behind in the AI development but also must rely on foreign systems and expertise for digital infrastructure.
Many regions of the Global South are rich in data owing to the accelerated rate of mobile penetration, increasing number of internet users, agricultural sensors, and health records. However, these regions lack robust data governing mechanisms. The data of the Global South often ends up being extracted and monetised by foreign platforms for their own gains. The sheer scale of data offered by the countries in the Global South massive. For example, over 1.3 billion people have registered in India’s biometric identification system Adhaar. Similarly, over 228 million people have registered into Pakistan’s National Database and Registration Authority (NADRA). However, in absence of indigenous data processing and protection mechanisms, the data of Global South ends up flowing to the servers of private corporations and companies of the developed world such as Google, Microsoft, Meta, Huawei, Amazon and Tencent.
Next is the asymmetry in the governance frameworks of AI. Many countries of the world have already rolled out their AI strategies and plans. However, the leadership role remains reserved by a few including the United States (US), China, France, and UK. In Europe, EU AI Act (2024) continues to offer defining framework for development and regulation of AI across the region and beyond. In Global South, however, the AI strategies often focus on ambitious goals without clearly defining binding regulatory mechanisms, enforcement architecture, budgetary commitments, and institutional mandates. Moreover, governments in these countries often rely on the foreign vendors and international consultants for expertise. Thus, these governments end up setting and enforcing digital norms that do not reflect local realities.
Lastly, one of the most fundamental asymmetries can be observed in the domain of digital literacy between the developed countries and the Global South. According to latest estimates, around 93 percent of the population in the high-income countries use internet as compared to 38 percent in Africa and 52 percent in South Asia which clearly shows the asymmetry in digital connectivity and access to AI tools. On top of that, only 5 percent of the people on low-income countries and 15 percent in the lower-middle income countries possess basic digital skills. Low levels of digital literacy increase the risks of digital manipulations, disinformation and online extremism in the Global South.
Global South cannot simply afford to engage in ‘botlash’ as it is a product of AI saturation and development, which is yet to happen for the Global South. Without localised research, development and governance of AI, many countries in the Global South could potentially become the testing grounds for new technologies. To avoid this trap, developing countries must invest in enhancing their digital literacy rate and build domestic capacity of AI R&D. Moreover, there is also a need to establish independent AI oversight and governing bodies and to prioritise local language and culturally relevant models of AI.
Muhammad Faizan Fakhar is a Senior Research Associate at the Centre for Aerospace & Security Studies (CASS), Islamabad, Pakistan. The article was first published in South China Morning Post. He can be reached at: [email protected]
Share this article
The Cover-up: IAF Narrative of the May 2025 Air Battle
Even after one year since the India-Pakistan May war of 2025, the Indian discourse regarding Operation Sindoor remains uncertain under its pretence of restraint. The Pahalgam attack on 22 April, which killed 26 people, triggered an escalatory spiral. New Delhi quickly accused Pakistan-linked elements, while Islamabad refuted the allegation and demanded an independent investigation. On 7 May, India launched attacks deep inside Pakistan under what it later termed as Operation Sindoor. The political motive was intended to turn the crisis into coercive signalling by shifting the blame onto the enemy and projecting a sense of military superiority.
This episode, however, began to fray immediately as war seldom follows the intended script. Within minutes PAF shot down 7 IAF aircraft including 4 Rafales. On 8 May, Reuters reported that at least two Indian aircraft were shot down by a Pakistani J-10C, while the local government sources reported other aircraft crashes in Indian-occupied Jammu and Kashmir
Why the IAF’s Post-Sindoor Spending Surge is a Sign of Panic
After Operation Sindoor, India is spending billions of dollars on new weapons. This is being taken by many people as an indication of military prowess. It is not. This rush to procure weapons is in fact an acknowledgement that the Air Force in India had failed to do what it was meant to do. The costly jets and missiles that India had purchased over the years failed to yield the promised results.
Sindoor was soon followed by India in sealing the gaps which the operation had exposed. It was reported that Indian Air Force (IAF) is looking to speed up its purchases of more than 7 billion USD. This will involve other Rafale fighter jets with India already ordering 26 more Rafales to the Navy in 2024 at an estimated cost of about 3.9 billion USD. India is also seeking long-range standoff missiles, Israeli loitering munitions and increased drone capabilities. Special financial powers of the Indian military were activated to issue emergency procurement orders. The magnitude and rate of these purchases speak volumes.
Indian media and defence analysts have over the years considered the Rafale as a game changer. When India purchased 36 Rafales aircrafts at an approximate cost of 8.7 billion USD, analysts vowed that the aircraft would provide India with air superiority over Pakistan. Operation Sindoor disproved all those allegations. Indian aircraft did not even fly in Pakistani airspace when the fighting started. India solely depended on standoff weapons that were launched at a safe distance. The air defence system of Pakistan, comprising of the HQ-9 surface-to-air missile system and its own fighters, stood its ground.
May 2025: Mosaic Warfare and the Myth of Centralised Air Power
Visualise a modern-day Air Force commander sitting in the operations room, miles away from the combat zone, overseeing every friendly and enemy aircraft and all assets involved in the campaign. In a split second, he can task a fighter, reposition a drone, and authorise a strike. In today’s promising technological era, he does not even need an operations room; a laptop on his desktop will suffice. The situation looks promising as it offers efficiency, precision, and control. The term used for such operational control is ‘centralisation’, which has been made possible with advanced networking, integrating space, cyber, surveillance, artificial intelligence, and seamless communication, enabling a single commander to manage an entire campaign from a single node. Centralised command and control, championed by the Western air forces and then adopted by many others, has thus been seen as a pinnacle of modern military power.
The concept of centralisation, enabled by state-of-the-art networking, may seem promising, but it is nothing more than a myth.
