Core takeaway: Despite 73% of humanity now online, 2.2 billion remain disconnected, and the AI age is not closing this gap—it is deepening it. Generative AI adoption in developed countries nearly doubles that in developing nations, threatening to fracture the world into a K-shaped society where a few build AI systems and the rest inhabit spaces designed by others.
The Persistence of the Old Divide
The most basic layer of the digital divide is access. In high-income economies, internet use reaches 94 percent of the population. In low-income countries, that figure plummets to just 23 percent. The International Telecommunication Union‘s 2026 ICT Development Index, covering 159 economies, shows an average score of 42 in low-income economies compared to 93 in high-income ones — a 75 -point gap that underscores persistent disparities in affordability and internet use.
Geography compounds the problem. While 86.5 percent of urban residents can access the internet, only 54.5 percent of rural inhabitants can. Of the 2.6 billion people offline, 1.8 billion live in rural areas. Even where networks exist, adoption lags. Mobile broadband now covers over 90 percent of the global population, yet in countries like Uganda and Zambia, only 9 percent and 17 percent respectively actually use the internet.
The barriers are not just infrastructure. They are affordability, skills, and trust. Four times as many people own smartphones in high-income countries compared to low-income ones. In low- and middle-income countries, 810 million women remain offline — 200 million fewer women connected than men. And as the World Bank notes, in 27 countries, more than half the population has never even sent a text message.
The AI Divide: A New Fault Line
The AI age is not waiting for the digital divide to close. In the first quarter of 2026, 27.5 percent of working-age people in developed countries used generative AI tools, compared to just 15.4 percent in the developing world — and that gap widened by 1.5 percentage points from late 2025. The Global North is adopting AI at nearly twice the rate of the Global South.
This is not merely a technology gap. It is an agency gap. As UNICEF warned in January 2026, “AI is advancing with limited regard for people already left behind, widening inequalities along familiar lines of income, geography and opportunity”. The dangers to social justice are greatest where AI-shaped decisions affect those who rely most on public services and are more likely to be digitally excluded due to age, poverty, or disability.
The compute gap is even starker. Sub-Saharan Africa faces a 98 percent compute gap; South Asia faces an 87 percent gap. The United Nations’ top technology official warned that the world risks fracturing into a “K-shaped society” where a handful of actors architect AI systems while the rest inhabit spaces designed by others. As he put it: “There’s the digital divide. Now there’s the data divide. Now you have a huge AI divide”.
Who Gets Left Behind?
The answer is not a single group. It is the intersection of multiple disadvantages:
• The poor: In low-income countries, where three-quarters remain offline, AI-driven economic opportunities are simply out of reach.
• The elderly: In many countries, older adults are the least likely to adopt digital tools — and the most likely to be excluded from AI-mediated services.
• Women: The gender gap in mobile internet access persists, with women in LMICs 12 percent less likely to use mobile internet than men.
• Rural communities: With 54.5 percent rural internet access versus 86.5 percent urban, the infrastructure gap remains a chasm.
• Children and youth in low-income households: As one report noted, children from lower-income households are less likely to discuss AI at school, reflecting wider inequalities in media literacy education.
The Cost of Inaction
The consequences are compounding. Countries that lag in connectivity risk falling further behind in productivity, innovation, and human development. AI does not just answer questions — as UNICEF observed, “It quietly teaches children what is possible for them”. When some children find AI tutors that adapt to their pace and build their confidence, while others are met with warnings that the service is unavailable or requires payment, the gap in capability and aspiration widens from the earliest age.
ITU chief Doreen Bogdan-Martin described the current moment as “a tale of two digital realities between high-income and low-income countries”. Cosmas Luckyson Zavazava, head of the ITU’s Telecommunication Development Bureau, put it more bluntly: “The world is inching towards universal access at a time that it should be sprinting”.
The AI age will not wait. The question is not whether AI will reshape the world — it already is. The question is whether that reshaping will lift everyone, or whether billions will be left behind, not just offline, but invisible in the datasets that train the systems of tomorrow. Bridging the digital divide in 2026 is no longer about catch-up. It is about preventing a permanent split.
