A yawning gap separates artificial intelligence’s promise in Europe and Central Asia from the region’s capacity to deliver on it. Fewer than one in ten companies there currently deploy AI, a figure roughly half the European Union average.
Productivity growth has stagnated across the region for two decades. Working-age populations have already begun shrinking in most countries, with much steeper declines projected by 2060. Meanwhile, education quality deteriorated over the past decade in most ECA nations. As a result, the latest World Bank Economic Update warns that missed opportunities loom large.
What blocks wider AI adoption? Firms consistently point to people and skills as the primary constraint, not access to advanced models or computing power. Most businesses now using AI apply it superficially, generating little real productivity gain. The exceptions prove instructive. Banks screening loan applications, retailers coordinating stock across suppliers, and energy producers optimizing inputs all invested heavily in workforce capabilities, data infrastructure, and results-focused management.
Four elements distinguish successful adopters. First, workers need solid literacy, numeracy, and reasoning foundations. A significant portion of ECA’s labor force lacks these basics. Second, vocational education must deliver relevant competencies. The region enrolls 45 percent of upper-secondary students in vocational tracks, more than double the global average. Third, managers need the ability to redesign production processes. Research suggests management practices explain roughly a quarter of productivity gaps between and within countries. Fourth, continuous learning matters. Employers and governments must support workers who keep upgrading their skills as labor markets shift.
PISA 2025 results compound the concern. Average scores across math, reading, and science slipped from 421 points in 2018 to 406 in 2025, equivalent to losing nearly a year of schooling. In most ECA countries, at least half of 15-year-olds score below minimum proficiency. In Kosovo, Armenia, Kyrgyzstan, and Tajikistan, that share exceeds four in five.
Critical thinking gaps prove especially damaging. Students who verify AI outputs and apply source-checking score about three years ahead of those who accept results at face value. Only 29 percent of ECA students take that careful approach, versus 45 percent in the EU. Using AI merely to summarize texts yields no measurable learning benefit.
Some countries show reversals are possible. Türkiye, Montenegro, Georgia, and Uzbekistan rank among the six largest PISA improvers worldwide. Türkiye paired digital infrastructure with curriculum reform shifting away from rote memorization. Uzbekistan set a national target for PISA top 30, banned phones in schools, and expanded early childhood education from 30 to 75 percent coverage.
Education reform now determines whether AI amplifies thinking or substitutes for it. Without foundational skills, the region risks splitting into a small elite capturing AI gains while the majority falls further behind.
