AI is already used by nearly one fifth of working-age people. The gap between richer and poorer countries persists

Microsoft released another Global AI Diffusion Report, which attempts to measure the actual spread of generative AI across countries. According to its estimate, 18.8 percent of the world’s population aged 15 to 64 used some generative-AI product in June 2026. Compared with the first quarter, that is an increase of roughly one percentage point. The methodology is based on aggregated and anonymized operational data from Microsoft. The company adjusts the figures for the share of operating systems and devices, internet connectivity and population size. Microsoft itself notes that this is not a perfect measure. That is an important caveat. Telemetry from one technology provider cannot precisely capture the entire global market. In some countries, other platforms and local models hold dominant positions. The result is therefore more useful for tracking the trend than as an exact figure for each country. The trend is nevertheless clear. AI use is growing in almost all economies. Generative models have reached a level of adoption within a few years that took previous enterprise technologies substantially longer. Differences between countries persist, however. Access to AI is not merely a question of whether a website is available. It requires good internet access, devices, digital skills and, for paid services, sufficient purchasing power. Microsoft also points to a different structure of use. In countries of the Global South, education and learning account for a larger share of activity than in richer economies. This could have a significant development effect: AI can provide inexpensive access to explanations, translation and individualized help where teachers or specialist materials are scarce. At the same time, a new digital divide may emerge. If the most capable models, high-quality data infrastructure and agent tools are available mainly in richer economies, a free basic chatbot may not narrow the productivity gap. For companies, the 18.8 percent figure is also relevant to customer behavior. If a significant part of the population begins to search for information through conversational models, the distribution of digital content, e-commerce and customer support will change. Pressure on the labor market is also increasing. Some employees use AI without formal deployment by their employer. Organizations that merely ban use may gain less control, not less usage. A practical response is to provide approved tools and rules for handling sensitive data. For governments, infrastructure and education will be crucial. Subsidizing access to a chatbot alone has limited effect if people do not know how to verify the output or lack digital services into which AI can be meaningfully integrated. The next stage of measurement therefore needs to track not only the number of users but also actual economic outcomes. Using AI once a month is different from changing a work process every day. KEY TERMS AI diffusion: The spread of technology use among people, companies and economies. Telemetry: Aggregated operational data collected from the use of digital services. Digital divide: A gap in access to digital technologies and in the ability to use them effectively.

AI is already used by nearly one fifth of working-age people. The gap between richer and poorer countries persists

The full article is available to users registered at Hard Skills