Global AI Usage Grows, but Not Evenly
New data from Microsoft’s Global AI Diffusion Report shows global AI usage continuing its upward trajectory. In the first quarter of 2026, the share of the world’s working-age population using generative AI rose by 1.5 percentage points, from 16.3% to 17.8%. That shift suggests the market is progressing beyond early experimentation into broader, everyday use. Twenty-six economies now report more than 30% of people aged 15 to 64 using AI tools, a sign of deepening integration into work and study. Yet this growth masks sharp regional disparities. While some countries are racing ahead, others are only inching forward, revealing a structural AI adoption gap. Microsoft’s methodology, based on anonymized telemetry adjusted for device mix, internet access, and population, underscores that AI diffusion is not just a function of product availability but of broader digital readiness.
A Widening AI Skills Divide Between Regions
The latest figures highlight a clear AI adoption gap between high-income and lower-income regions. In the Global North, 27.5% of the population used generative AI in Q1, up from 24.7% in the second half of 2025. In the Global South, usage rose more modestly, from 14.1% to 15.4%. The difference between these blocs has therefore widened from 10.6 to 12.1 percentage points. This AI skills divide reflects uneven diffusion of tools and capabilities: some economies are rapidly embedding AI into professional workflows, while others struggle to build basic familiarity. Microsoft’s data suggests the Global North is not just adopting AI faster, but also increasing intensity of use. As more economies surpass the 30% usage threshold, the risk grows that lagging regions will fall further behind in productivity, innovation capacity, and access to emerging AI-enabled jobs.
Infrastructure, Literacy and Access Shape Emerging Markets AI
Microsoft links the uneven spread of AI to foundational differences in infrastructure and human capital. Reliable electricity, consistent internet connectivity, and baseline digital skills all influence whether people can meaningfully use generative AI. In many emerging markets, these prerequisites remain patchy, limiting both adoption and depth of use. The issue, therefore, extends beyond simply making AI tools available or translating interfaces. It becomes an education and workforce challenge: equipping students, teachers, and workers with the competencies needed to integrate AI into everyday tasks. Without targeted investments in connectivity, devices, and digital literacy, emerging markets AI adoption risks stagnating. That would entrench a two-speed global economy in which some societies leverage AI to accelerate innovation and others remain largely spectators, dependent on imported technologies and external expertise.
Usage Patterns: Downloads Surge, Activity and Retention Lag
While AI apps are increasingly visible in mainstream digital life—reaching top 10 download rankings globally in April—usage patterns remain volatile. Initial enthusiasm often translates into quick installs, but sustained engagement is harder to achieve. Many users experiment briefly with generative AI tools before reverting to familiar workflows, producing declines in user activity and inconsistent retention. This dynamic is particularly pronounced in markets where digital literacy is uneven: new users may lack clear use cases, training, or workplace support to integrate AI into daily routines. As a result, headline metrics for global AI usage can rise even as individual usage intensity plateaus or drops. For policymakers and companies, this underscores that closing the AI adoption gap is not just about access, but about sustained, purposeful use that turns experimentation into productivity gains.
Language, Coding and the Next Phase of AI Diffusion
Regional trends also show how product design can reshape the AI adoption landscape. In Asia, twelve of the fifteen fastest-growing AI adopters since mid-2025 are in the region, with South Korea, Thailand, and Japan recording the largest gains. Improved support for local languages and multimodal interaction has made AI tools more useful for search, learning, messaging, and content creation. In Japan, for instance, model performance on professional exams and benchmarks has risen sharply, helping drive adoption beyond early adopters. At the same time, AI coding platforms such as GitHub Copilot and other tools from OpenAI and Anthropic are boosting software development activity, with Git pushes and new repositories increasing significantly. These developments suggest that where infrastructure and skills are in place, AI can rapidly embed itself into both knowledge work and technical fields, further widening the gap with regions still building basic digital capacity.