I use artificial intelligence on regular basis, for a variety of reasons: coding apps and websites on demand, generating pictures and videos (mostly for fun), searching for information, translating articles from languages I don’t speak, and others. In a recent article for Medium, I argued why we should embrace AI rather than complain about it.
AI is a diversity of tools and tools need to be mastered, not destroyed. I mostly use the free versions of AI apps; only in two cases I buy subscriptions — for coding and image generation. So, my degree of AI adoption is perhaps around 90%.
Four years ago, when the first advanced generative AIs were released for the public, the world had mixed feelings about them — some saw opportunities, others saw existential threats to their livelihoods.
During this time framework, everyone realized that AI is an economic necessity that provides competitive advantage in every field — defense, energy, technology, finance, to name a few.
As a European, I wondered how many people on the old continent use AI for work or leisure — and where AI is most popular. So, I browsed the E.U.’s statistical agency Eurostat for data.
The findings are revelatory but also predictable.
Credit: Unsplash
Europe is a mosaic in almost everything — political systems, economic models, social welfare, languages, cultures, religions, and so on. AI is not an exception.
Right now, it is splintering into three distinct technological blocs. In Northern countries, AI has rapidly become a mainstream tool used by nearly half the population. In the West, adoption is slower but steady. In the South and East, adoption remains stagnant, with fewer than one in five people engaging with the technology.
It’s not about infrastructure or wealth. A closer scrutiny of the past two years reveals a powerful correlation between a nation’s literacy rates and its citizens’ willingness to embrace artificial intelligence. As AI becomes less about novelty and more about utility, the foundational skill of reading comprehension is emerging as the new determinant of digital inclusion.
A three-speed continent
According to Eurostat data published in April 2026 for Q4 of 2025, the disparity in AI adoption across Europe is stark. Northern Europe dominates the rankings, with 56.32% of individuals in Norway reporting use of generative AI in the last three months, followed by Denmark at 48.44%, Estonia at 46.44%, and Finland at 46.27%.
In stark contrast, adoption plummets in Southeastern Europe. Romania ranks last in the E.U. at just 17.76%, while Italy (19.86%). Among non-E.U. countries, Türkiye (17.195) and Bosnia-Herzegovina (20.26%) also languish near the bottom. Major economic powers sit awkwardly in the middle: Germany reported usage rate of 32.25%, France reported 37.46%, and Spain 37.88%.
The United Kingdom reported approximately 45% in December 2025.
This data suggests that economic output alone does not drive AI usage. If GDP were the primary factor, Germany would be leading the pack. Instead, the data points toward a more subtle, human-centric variable: proficiency in literacy and problem-solving.
Use of generative AI tools in the last 3 months of 2025; for private purposes; for work; for education. Source: Statista
No data exists on Moldova, Ukraine, Belarus and Russia, which are also European countries.
The literacy-AI correlation
To understand this gap, we turn to the OECD’s Survey of Adult Skills (PIAAC), which measures literacy proficiency among adults aged 16–65. The scores are measured on a 500-point scale and categorized into levels from “Below Level 1” (basic comprehension) to Level 4/5 (complex reasoning). The average OECD literacy score is 259 points.
When we overlay the AI adoption map with the OECD literacy rankings, the pattern becomes undeniable.
Countries leading in AI — Norway, Finland, Denmark, and Sweden — are also the literacy capitals of the world. Finland consistently scores at the top with an average literacy score of 296 points, well above the OECD average. These nations possess a populace that is not just capable of reading, but of critically analyzing dense texts.
Generative AI requires users to write effective prompts and critically evaluate outputs. A population with high literacy is inherently better equipped to navigate the probabilistic nature of AI.
Italy, Portugal, and several Eastern European nations struggle with both AI adoption and adult literacy. The OECD notes that in countries like Poland and Portugal, a significant portion of adults perform at or below Level 1 proficiency, meaning they struggle with lengthy or abstract texts.
In Romania and Bulgaria, lower scores in international assessments correlate directly with the lowest AI usage figures in the Eurostat survey.
Why literacy is the gatekeeper
What is the connection between an individual’s reading skill and interaction with a chatbot? The answer lies in the nature of the technology itself. Just a few thoughts:
1. Using a tool like ChatGPT is not a passive activity. Effective users must articulate a goal, provide context, and set constraints. This requires the ability to synthesize information and write clearly — skills taught through literacy education.
2. Critical validation: Generative AI produces “hallucinations” (incorrect information). A literate user can fact-check and identify logical inconsistencies. A user with low literacy may accept false information as truth, making AI a liability rather than an asset.
3. Relevance. The primary reason non-users avoid AI is that they don’t see the tools as relevant. This perception of irrelevance is often tied to digital literacy; if a user struggles to see how text-based tools apply to their low-text workflow, they will not adopt it.
The youth factor
There is a caveat to the “literacy determines adoption” thesis: the generational shift.
Across all European nations, AI usage among 16 to 24-year-olds averages 64%, compared to the 7% usage among those over 65.
Younger cohorts in Southern Europe are likely bridging the literacy gap. A 20-year-old in Italy or Türkiye may have higher digital and textual literacy than their parents, suggesting that while the current national map shows a North-West-South divide, the future may see a convergence as Generation Z enters the workforce.
Integration of AI in academic curricula is therefore crucial for AI adoption at national level. If you leave school at 18 and never engage with a large language model, you are unlikely to start using it at 45.