AI Readiness in Low- and Middle-Income Countries: Key Building Blocks

Artificial Intelligence (AI) has sparked enthusiasm for its potential role in development in Low- and Middle-Income Countries (LMICs) (AI Safety Summit, UK Gov, 2023). This excitement is grounded in the historical role that technological innovation has played in driving productivity, improving living standards, and enabling economic growth (OECD, 2015)—and in AI's emerging role as a general-purpose technology with broad cross-sector potential (Brynjolfsson, 2022).

This blog explores the idea that investing in foundational digital capacity, such as digital skills, internet access, mobile device penetration, and reliable electricity, could be one of the most effective and scalable ways to support both economic development and global AI inclusion. Rather than focusing on advanced applications or AI research centers alone, readiness should be built from the ground up.

Innovation Isn’t Automatic

Innovation, broadly defined as the introduction and adaptation of new products, technologies, or business models, is widely recognized as a core driver of long-term prosperity. However, as Cirera and Maloney’s Innovation Paradox illustrates, the presence of new technologies does not guarantee progress. Countries often fail to benefit from innovation unless they have the necessary supporting conditions—what economists call "complementarities."

Applying this lens to AI, the logic is straightforward: without digital foundations, AI investments may yield low or even negative returns. Countries lacking basic infrastructure, digital literacy, and institutional capacity are less likely to adopt, adapt, or develop AI tools that deliver value locally.

Digital Divide and Development Gaps

The World Bank, UNESCO, and the International Telecommunication Union (ITU) consistently highlight the digital divide—the gap in access to digital technologies—as a structural barrier to inclusive development. These institutions also underscore the importance of building the capabilities needed to benefit from these technologies. As the World Bank emphasized in its 2016 World Development Report, digital technologies alone do not deliver development gains; they must be matched with strong “analog complements”—such as effective institutions, skills, and regulations. Without these, technology risks amplifying existing inequalities rather than closing them. Meanwhile, UNESCO and ITU note in their State of Broadband 2024 report that “broadband and computing infrastructure provides the basis for changes brought about by AI.”

Internet access is a basic requirement for any AI interaction. In lower-middle-income countries, only 57% of the population is online (ITU, 2024). In Sub-Saharan Africa, the figure drops to 41%, the lowest among all regions. However, progress is notable: between 2016 and 2023, Sub-Saharan Africa saw internet usage more than double. Understanding and scaling the drivers behind this trend will be crucial for continued digital expansion, particularly as compounding effects such as network externalities and lower per-user costs make digital infrastructure more valuable with scale.

Coverage improvements are especially striking. According to ITU statistics, LTE/WiMAX (4G) coverage across Africa—measured by the share of the population living within reach of a signal—grew more than sixfold between 2015 and 2024, rising from just 11% to over 70%. This represents a remarkable expansion in population-level access to mobile broadband and enables broader reach of basic mobile-based AI tools. Yet gaps remain: 5G coverage, critical for advanced, real-time applications, still reaches only 11% of the population in Africa, compared to over 70% in Europe. Moreover, access doesn’t guarantee usage. Yet as the World Bank (2023) notes, “84 percent [of people in Sub-Saharan Africa] live in areas where mobile internet services are available, yet only 22 percent used them by the end of 2021,” underscoring that coverage alone is not enough—affordability, relevance, and digital skills are equally critical for meaningful digital participation.

Other digital readiness indicators show similar patterns. Mobile phone usage in Africa rose from 54% in 2019 to 66% in 2024—an increase of 12 percentage points, or roughly 22% growth—yet still lags behind Asia-Pacific (77%) (ITU, 2024). And beneath these averages lie wide disparities, especially between urban and rural populations.

Even more foundational is electricity. According to UNESCO 2023, one in four primary schools globally lacks access to electricity, rising to over 70% in some rural parts of Sub-Saharan Africa. In schools, early digital exposure can help shape future skills, and while data across LMICs is limited, Rwanda and Senegal show notable progress. In Rwanda, the share of upper secondary schools with computers for pedagogical purposes rose from 69% in 2011 to 88% in 2023, and in Senegal from 85% in 2019 to 90% in 2022.

Internet access in secondary schools similarly improved in these positive case studies: from 17% to 79% in Rwanda (2011–2023), and from 46% to 68% in Senegal (2019–2022). These upward trends point to promising momentum, but also highlight the need for further research into how education systems in LMICs move from basic device access toward meaningful digital integration.

The Capabilities Escalator

Building on Cirera and Maloney's "capabilities escalator" framework, we can think of AI readiness not as a switch to be flipped, but a ladder to climb. Countries must first absorb technologies, then adapt them, and ultimately innovate upon them.

This ties closely to the concept of absorptive capacity—a country’s ability to recognize, assimilate, and apply new knowledge (Cohen & Levinthal, 1990; Danquah, Ouattara & Quartey, 2018). Without foundational systems in place, deploying advanced AI tools is unlikely to deliver meaningful or sustainable results. For example, AI diagnostics require digital health records, connectivity, and trained personnel. Similarly, students and entrepreneurs need internet-connected devices to access even the most universal AI tools.

A Framework for Readiness: The Six Capitals

To help structure where investment should be targeted, we draw on a "Six Capitals" framework. These are productive assets—both tangible and intangible—that underpin a country’s ability to benefit from digital innovation. Crucially, these capitals are interdependent: strengthening one often relies on progress in others. For instance, building human capital requires physical infrastructure like electricity and internet; deploying digital tools depends on institutional capacity and public trust; and producing and using high-quality data requires both technical skills and governance frameworks.

1. Human Capital: Digital literacy, technical skills, and managerial capacity are essential. A workforce that understands and can apply digital tools is better positioned to generate value.

2. Knowledge & Information Capital: Data is not just a byproduct of digital systems; it's a key input. High-quality, context-specific data enables AI tools to work effectively. Without it, AI models risk misrepresenting local realities.

3. Physical Capital: Infrastructure like smartphones, laptops, mobile towers, and electricity grids form the backbone of digital engagement. Without these, even the most universal tools remain inaccessible.

4. Institutional Capital: Effective regulations, public agencies, and coordination mechanisms shape how technologies are introduced and governed. Weak institutions can limit both adoption and trust.

5. Financial Capital: Access to public, private, and external finance enables investment in infrastructure and services. Without funding, even well-designed plans remain unrealized.

6. Social Capital: Trust in institutions, openness to innovation, and community networks support technology adoption. In many cases, the willingness to share data or experiment with new tools depends on these softer but critical foundations.

Implications for Policymakers and Funders

While public and philanthropic funding plays a critical role, much foundational digital infrastructure, particularly in connectivity and energy, will ultimately depend on mobilizing private capital, often through blended finance or public-private partnerships, and leveraging local capabilities, markets, and resources.

For those interested in AI for development, safety, and global inclusion, there appears to be a compelling case for prioritizing foundational investments. While interest in frontier AI labs and advanced applications in LMICs is rising, the most cost-effective opportunities may lie in enabling meaningful participation in the digital economy by scaling access to connectivity, computing devices, and basic digital skills.

Evidence from The Innovation Paradox suggests that returns on innovation can exceed 100% in environments with strong foundations, and fall to zero or negative in low-capacity settings. That insight should inform not only government strategies, but also philanthropic and donor priorities.

Investing in digital infrastructure—from internet access to national data protection laws—can be a critical enabler of broader AI readiness. These efforts won’t just expand access to tools like large language models or mobile diagnostics; they can also support more inclusive and resilient governance of emerging technologies.

A Closing Thought

While AI has understandably risen on the global development agenda, this should not come at the expense of the longer-standing, unfinished work of digitalization. In fact, the two are inseparable. For policymakers and funders seeking high-impact, scalable interventions, foundational digital capacity appears as a highly promising lever, particularly in terms of long-term development impact and policy cost-effectiveness.

This initial data exploration highlights broad trends but simplifies the wide variation across and within countries. Further research is needed to identify which foundational investments deliver the greatest returns in different national contexts and how they align with broader development goals. This would help target resources where they can have the most effective and equitable impact—especially as AI raises both the stakes and the potential benefits.