Kenya has the world’s highest rate of AI use, with 97.5% of online adults reporting they used at least one AI tool in the past month, according to the Digital 2026 Mid-Year Global Update published by DataReportal in partnership with We Are Social and Manochi.
That is the highest figure among the 54 countries tracked — ahead of the United Arab Emirates at 94.2% and Indonesia at 93.6%, and far above major Western economies. Just 71.5% of online adults in the United States reported using AI in the past month, nearly 10 percentage points below the global average. The UK stood at 62.8%. Japan was the only market surveyed where fewer than half of respondents reported any AI use.
The figures come from GWI’s Core survey, which polled more than 240,000 people aged 16 and above across the world’s largest economies — markets accounting for over 80% of the global internet population.
Reconciling 97.5% with 8.7%
Three months ago, Microsoft’s Global AI Diffusion report placed Kenya at 8.7%, near the bottom of its global table and well behind South Africa’s 19%. Both figures are from 2026. Both are credible. They measure different things.
DataReportal surveys online adults and counts any AI tool used monthly, including capabilities embedded in software people may not think of as AI at all — Microsoft Office, Google Workspace, Canva — alongside standalone platforms like ChatGPT. Microsoft derives its estimate from aggregated product telemetry across the whole population, adjusted for operating system and device market share, internet penetration and population, and focuses on generative AI specifically.
The population base explains most of the gap. Kenya’s internet penetration sits at roughly 40%, so a figure covering online adults describes a minority of Kenyans. A whole-population measure necessarily produces a much lower number for the same underlying behaviour.
The two findings are therefore compatible, and together they say something more precise than either does alone: among Kenyans who are online, AI use is close to universal — but most Kenyans are not online.
What the numbers do and don’t show
DataReportal’s authors are explicit that adoption measures usage rather than depth of engagement. Finding information remains the top reason people worldwide turn to AI tools, ahead of seeking advice or improving skills.
That caveat aligns with PwC’s survey of nearly 50,000 workers, including 1,753 across Algeria, Kenya, Morocco, Nigeria and South Africa, which found Africa leading globally on overall workplace AI adoption while only 17% of African workers used AI agents daily — with most not using AI for workflow automation or advanced task support.
Globally, the report estimates 4.02 billion adults — 48.6% of the world’s population — now use some form of AI at least monthly. Active use of generative AI platforms has more than doubled over the past year, growing 141% to 2.42 billion users. The authors note that users in lower- and middle-income countries tend to use AI more frequently than the global average, suggesting adoption in unsampled markets could be higher still.
The policy side is catching up
Kenya’s usage lead comes as the country moves to formalise the sector. Public consultation recently closed on the draft Kenya Artificial Intelligence and Other Emerging Technologies Policy 2026, developed by a technical working group drawn from academia, government, startups and civil society, with support from the Kenya ICT Action Network, the British High Commission, the European Union, Germany and Canada.
Building on the National AI Strategy 2025-2030, the draft sets out a four-tier governance structure anchored by a National AI and Other Emerging Technologies Steering Committee and a new regulatory council. It proposes a Kenya AI Safety Institute for testing and risk classification, a public registry, and a National African Languages and Voice Programme aimed at building systems that work in local languages.
The policy is expected to move toward Cabinet consideration alongside a separate AI Bill currently before Parliament.
That last element is the one to watch. Near-universal AI use among online Kenyans, running overwhelmingly on tools built elsewhere and trained on data from elsewhere, is precisely the adoption pattern that makes a local-language programme consequential rather than symbolic.
