On January 28, 2026, the OECD published its latest AI adoption figures, drawn from its ICT Access and Usage Database: more than a third of individuals — 36.8 percent, as cited by ITIF — used generative AI tools in 2025, while firm-level adoption climbed from 8.7% in 2023 to 20.2% in 2025, doubling in two years. This is the broadest, most methodologically consistent cross-national benchmark available. While most AI discourse is still about model capabilities, this dataset hands product and policy teams rare hard evidence about who is actually using these tools — and who has not started.
Individuals: More Than a Third Already Use It
- 36.8% of individuals used generative AI tools such as ChatGPT or Claude at least once in 2025.
- Nordic countries lead: Denmark, Finland, and Sweden all sit above 35%.
- Status matters: three-quarters of students aged 16 and over use these tools; 41.1% of employed people do, 36.7% of the unemployed, and only 12.5% of retirees and others outside the labor force.
“One in three” is a psychological threshold: generative AI has crossed from an early-adopter novelty into mainstream everyday software. The three-quarters figure among students is a preview of the default skill set of the incoming workforce over the next five years. The other end of the distribution is just as telling: at 12.5%, retirees and people outside the labor force are the clearest marker of how far diffusion still has to go beyond work and school contexts.
The Shape of the Divide: Age, Education, Income
- Age is the steepest gradient: a 53.6 percentage-point gap between the most and least active age groups.
- Education and income each account for gaps of roughly 21 percentage points.
- Gender is comparatively small at 4.2 percentage points.
Generative AI is diffusing as an “age-first” technology — a sharper divide than the PC or smartphone eras. A 4.2-point gender gap is narrow by the standards of earlier technology waves; the defining axis of this one is generational. The implication for product teams is blunt: the friction that remains is concentrated in older and non-employed populations, and it is about interface and trust, not raw capability.
Firms: Doubled in Two Years, but Slowing
The firm adoption trajectory runs 8.7% (2023) to 14.2% (2024) to 20.2% (2025). Growth in 2025 came in at 42.4%, down from 59.6% in 2024 — still rapid diffusion, but decelerating as the base grows, exactly what the middle of an S-curve looks like. Even at that moderated pace, six percentage points of firms — on top of a 14.2% base — adopted during 2025 alone.
Put differently: individuals are past one-third, firms just past one-fifth. Organizations lag individuals, and that scissors gap is the market for internal tooling and adoption services.
Size and Sector: Who Is Leading
- Size gap is threefold: 52.0% of large firms use AI versus 17.4% of small firms.
- By sector, ICT leads at 57.3%, with professional, scientific, and technical services at 36.8%; among ICT firms, Sweden (87.9%), Austria (79.9%), and Finland (79.8%) are near saturation.
- The fastest growth sits in low-base traditional industries: accommodation and food services up 62.5%, construction up 59.1%.
Growth concentrating in hospitality and construction is worth a second look: these are frontline industries with thin margins and little in-house ML talent, which suggests the arriving use cases are customer communication, scheduling, and estimating rather than code. That is the profile of adoption driven by packaged tools, not by data teams.
Three Readings for Product Teams
First, in the consumer market the axis of competition has shifted from educating users to retention and depth — most potential users have touched generative AI at least once. Second, the enterprise gap lives in small firms and non-tech industries: at 20.2% overall adoption, most companies have not started, and for them ease of use and workflow integration beat model capability. Third, the 53.6-point age gap is a design signal: a product that serves only digital natives is walking away from nearly half its potential market. The dataset also backs what we argued in our 2026 opening outlook: this year’s story is depth of adoption, not model noise.
Sources
- AI use by individuals surges across the OECD as adoption by firms continues to expand — OECD
- 36.8 Percent of Individuals in OECD Countries Used Generative AI Tools in 2025 — ITIF
- Accelerated increase in the use of AI by individuals in OECD member countries — EU Digital Skills & Jobs Platform
AI-assisted summary compiled from the sources above, reviewed by a human before publishing.