17.8% of World's Workers Now Use AI, Microsoft Report Finds

Roughly one in six working-age adults worldwide now uses artificial intelligence tools on a regular basis, according to Microsoft's AI Economy Institute, whose latest quarterly diffusion report pegs Q1 2026 adoption at 17.8 percent — a 1.5 percentage-point increase over the final quarter of 2025.

The headline figure, drawn from the institute's National AI Leaderboard, tracks what the researchers define as consistent, work-related AI use among the 3.5 billion people of working age globally. It does not count one-off curiosity searches or incidental exposure; the methodology captures active, repeated engagement with AI tools in professional or productive contexts.

Who Leads and Who Lags

The United Arab Emirates holds first place on the leaderboard at 70.1 percent — a figure that reflects more than a decade of deliberate government investment in AI infrastructure, including a planned data center campus outside Abu Dhabi that will draw five gigawatts at peak demand and is being developed in partnership with US technology companies. Singapore, Norway, Ireland, and France round out the top five, a cluster that speaks to the role of small, high-income economies in early adoption.

Twenty-six economies now exceed the 30 percent threshold, up from fewer than twenty a year ago. The United States climbed three places to twenty-first with 31.3 percent, a ranking that may seem paradoxical given that Silicon Valley houses the largest concentration of AI developers in the world. The gap between building AI and using it at scale reflects how unevenly productivity tools diffuse through large, heterogeneous labor markets.

Asia showed the most pronounced movement within the quarter. South Korea, Thailand, and Japan all recorded notable upticks — driven, the report argues, by improvements in multilingual model quality that have finally made Japanese, Thai, and Korean speakers first-class citizens of large-language-model interfaces rather than second-tier users relying on machine-translated prompts. That shift, largely invisible in English-language technology coverage, could have outsized implications for adoption curves in the 800 million people whose primary working language is one of those three.

The Adoption Gap Widens

The report's most sobering finding is structural rather than celebratory. The Global North as a whole stands at 27.5 percent adoption; the Global South at 15.4 percent. The 12-point gap is wider than it was six months ago, not narrower, despite the falling cost of inference and the proliferation of free-tier consumer products.

The institute attributes the divergence to several compounding factors: lower broadband penetration, fewer native-language models for the 6,000-plus languages spoken in sub-Saharan Africa and South Asia, weaker institutional frameworks for AI policy, and lower per-capita device ownership. Falling API costs help organizations already embedded in the AI supply chain; they do not automatically reach a smallholder farmer in Nigeria or a garment worker in Bangladesh without deliberate distribution infrastructure.

This is not a new observation — the Stanford AI Index flagged the same dynamic in its 2026 edition in early May — but the Microsoft data adds quarterly granularity that allows comparison across time, and the Q1 trajectory is moving in the wrong direction on equity terms.

Software Development as a Leading Indicator

One data series in the report deserves particular attention from anyone tracking whether AI is displacing or augmenting technical workers. Git pushes globally increased 78 percent year-over-year, a proxy metric for how much code is being written and shipped. Simultaneously, US software developer employment reached approximately 2.2 million in 2025 — an 8.5 percent annual increase and a record high for the profession, with early 2026 figures running about four percent above March 2025 levels.

The institute's interpretation is that cheaper development costs are expanding the total amount of software being built rather than simply replacing the labor that built it before. More software gets commissioned when the marginal cost of each feature drops; more developers are hired to manage the expanding scope. That argument is plausible for the period immediately following a major capability transition, though it says little about the medium-term trajectory once organizations have fully internalized the new cost structure.

What to Watch

Three pressure points will determine whether the Q2 2026 figures continue the upward trend or plateau. The first is the wave of AI-native consumer and enterprise products expected from Google I/O on May 19, which may pull casual users into sustained engagement. The second is whether any major economy in the Global South breaks clearly above the 30 percent threshold, which would suggest the adoption gap is narrowing rather than structurally locked in. The third is the jobs data: if developer employment peaks while code output continues to rise, the labor-complementarity thesis will face its first serious empirical challenge.

Microsoft's AI Economy Institute publishes quarterly updates; the next is expected around the time of the firm's Q3 earnings in late July.