OECD: GenAI Chatbot Usage Has Grown From 18% to 28% of the Population in One Year

An analysis by the Organisation for Economic Co-operation and Development, published in late June 2026, presents the first comprehensive data on how residents of 37 countries in the Global Partnership on AI (GPAI) use generative AI chatbots — and the figures reveal an acceleration rarely seen in digital technologies.

Growth That Outpaces Previous Digital Waves

Data collected by SimilarWeb between February 2024 and March 2026 shows that the share of the population actively using GenAI chatbots rose from approximately 18% (January 2025) to 28% (January 2026) — an increase of ten percentage points in just one year. For comparison, smartphone penetration took almost a decade to achieve a similar jump in developed markets.

Market concentration is pronounced: ChatGPT, Claude, and Gemini together account for virtually all web traffic to chatbot interfaces across all surveyed GPAI countries, meaning the market is effectively a three-tier oligopoly.

Singapore, Japan, and Turkey Lead in Adoption

The geographic distribution of adoption is far from uniform. Singapore records a rate of 63% of the population using AI chatbots — nearly double the global average and a jump from 36% a year earlier. Japan and Turkey more than doubled their usage rates between 2024 and 2025, placing them alongside fast-growing markets that combine high digital literacy with active institutional incentives.

This geographic divergence points to differing combinations of factors: regulatory framework, digital literacy, infrastructure speed, and the presence of local AI initiatives in education and business.

Who Are the AI Chatbot Users?

The demographic profile clearly shows the dominance of younger professionals. More than 50% of the 25–34 age group used chatbots in February 2026. At the opposite end of the spectrum, people aged 65 and over use these tools in only 8% of cases — a gap of more than 40 percentage points within a single population.

It is worth noting that the “aging” of the user base is accelerating. The share of users over 35 has risen from 38% (February 2024) to 48% (February 2026) — meaning chatbots are slowly penetrating demographic segments that did not initially adopt them quickly.

Gender dynamics are more complex than intuition might suggest. Globally, men use chatbots somewhat more, but the gap is smaller than in many other technologies. The exception is South Korea, where usage among men is nearly double that among women. On the other hand, 10 countries — including Singapore, Colombia, the Czech Republic, Finland, Hungary, Ireland, Mexico, and New Zealand — record higher adoption rates among women.

User Behavior Differs Across Platforms

Analysis of web traffic patterns reveals an important difference in user profiles across platforms.

Claude and Microsoft Copilot users visit LinkedIn, GitHub, and Notion significantly more often — suggesting predominantly professional and development-oriented use. These users are looking for a tool that helps with professional work: writing code, analyzing documents, managing projects.

ChatGPT and Gemini users, by contrast, gravitate toward YouTube, Instagram, and Facebook, suggesting a broader, recreationally and professionally mixed base — from students to creatives looking for inspiration or quick answers.

Average visit duration is increasing across all platforms: from 4.5 minutes (mid-2024) to more than 5.5 minutes (early 2026), indicating ever deeper integration of chatbots into daily workflows rather than occasional queries.

Methodology Reliability and Data Limitations

The OECD conducted external validation by comparing SimilarWeb estimates with two independent sources. Correlation with Eurostat survey data stands at 0.63, and with data from the Microsoft AI Economy Institute at 0.78 — which the researchers consider a solid result for a web-traffic-based methodology, especially given differences in definitions and data collection periods.

Methodology limitations remain: SimilarWeb data does not cover mobile app usage or API integrations within business tools. Actual usage of GenAI systems is therefore likely higher than the percentages shown.

Implications for Policy and Industry

With this research, the OECD is sending a clear message to policymakers: the gap between older and younger users, as well as gender differences in certain countries, calls for targeted digital education interventions. The high adoption rates among younger generations — particularly in Singapore — suggest that AI literacy is becoming a critical labor-market competency on par with knowing how to use spreadsheet tools in the 1990s.

For industry, the key information is the growing average visit duration: users are not experimenting — they are integrating AI tools into regular work processes. This opens space for more specialized, professionally oriented services targeting segments that dominant platforms have yet to fully cover.