AI access gap spans countries and workplaces as adoption rises

The AI access gap is showing up at two different levels: Microsoft reported that the divide in generative-AI use between the Global North and Global South widened in the second half of 2025, while an Infosys survey found a separate difference in access to AI tools and training inside enterprises. The measures are not comparable, but together they show that a higher headline adoption rate does not establish evenly distributed access, training or organizational readiness.

Microsoft’s AI Economy Institute said global generative-AI use reached 16.3% of the world’s population in the second half of 2025, up from 15.1% in the first half. Its measure defines AI diffusion as the share of people who used a generative-AI product during the reporting period.

The company put usage among working-age people at 24.7% in the Global North and 14.1% in the Global South. It said adoption in the North grew nearly twice as fast, increasing the gap from 9.8 percentage points to 10.6 points. Microsoft’s figures are based on aggregated, anonymized telemetry, adjusted for operating-system and device-market share, internet penetration and country population.

That methodology provides a cross-country indicator, not a census of AI use. Microsoft said no one metric can fully capture adoption and that it expects to add indicators as its approach develops.

What does the AI access gap mean for companies?

Within companies, the problem is less about country-level diffusion than whether people doing the work have the tools, training and direction to use them. CIO Dive reported that the Infosys Knowledge Institute surveyed more than 2,600 white-collar workers for its AI and the Future of Work research.

More than one-third of senior leaders said employees were always equipped with the AI tools they needed, according to CIO Dive’s account of the survey. Only 19% of junior staff gave the equivalent response. CIO Dive also reported that Infosys found senior leaders had greater access to tools and training than regular employees.

The same research pointed to a strategic disconnect. Nearly half of senior leaders said their company intended to remake most or all work and processes with AI, while fewer than one-quarter of middle managers reported the same goal. Infosys characterized middle managers as central to both strategy and execution, able to shape how work is done and relay feedback on adoption plans.

For operators, the practical implication is to separate reported AI use from deployment quality. Daily usage can coexist with uneven permissions, weak training or a plan that is not understood below the executive level. Infosys reported that more than two-thirds of employees used AI at least daily and that users estimated an average time saving of four hours a week, but neither finding establishes consistent access across roles.

Microsoft identified early investment in digital infrastructure, AI skills programs and government adoption as common characteristics of leading countries, without presenting them as proven causes of the North-South difference. It also cited specific examples of accessibility affecting uptake: improved Korean-language model capabilities and consumer features in South Korea, and DeepSeek’s free chatbot and MIT-licensed open-source model reducing financial and technical barriers in some markets.

A later Microsoft update, as reported by Campus Technology, put working-age usage at 27.5% in the Global North and 15.4% in the Global South. The direction is clear in both snapshots, but the data does not settle how much of the divide is driven by infrastructure, affordability, language support, employer practices or other factors.

This story draws on original reporting from CIO Dive.