Artificial Intelligence and Structural Inequalities in Southern Europe: Comparative Evidence from Greece (2024–2026)
Christos Papagiannis
Adjunct Professor at Brussels School of Governance of Vrije Universiteit Brussel (VUB) and PhD Candidate at Maastricht University
Martin Wählisch
Associate Professor of Transformative Technologies, Innovation and Global Affairs at the University of Birmingham
This report draws upon research conducted by Eteron and AboutPeople (Nov 9-11 2024) and by diaNEOsis and Metron Analysis (January 8-13, 2026), examining whether the integration of AI over that two year period has had a positive or negative effect on social and economic inequalities. The first study was a CAWI (online interview) study with a nationally representative sample of 1,102 respondents. The second study used a mixed mode approach (CATI and CAWI) with a nationally representative sample of 1,111 respondents. The findings are supplemented by 2025 Eurostat data, comparative data from a number of Southern European countries and a number of insights from the Special Eurobarometer.
The approach of this report is descriptive and exploratory (rather than strictly longitudinal) as the studies in question vary in terms of methodology, the profile of the respondents and also in the way similar questions were expressed and asked. More specifically, it should be noted that the exclusively online methodology of 2024 likely reached respondents who were already more familiar with digital technologies, whereas the mixed-method approach of 2026 is more inclusive, as it can also reach citizens with little to no knowledge of the internet due to the mixed nature of its sample collection. The percentages derived therefore, should not be viewed as directly comparable, but rather as indicative evidence of a wider shift in which age, education, income, gender and position in the labour market all appear to be significant factors for AI diffusion. The research clearly concludes that we need policies to regulate AI equitably, given that the technology itself doesn’t seem to be an automatic equaliser. Therefore, the absence of tailored policies creates a risk of amplifying existing inequalities.
Specifically, the research notes the following:
- Greece is ahead of the EU in terms of individual usage. In 2025 44.1% of people in Greece had used generative AI in the previous 3 months compared to 32.7% in the EU-27.
- Private use is at 40.9% in Greece as opposed to 25.5% in the EU. Professional use is only just above the EU average at 16.1% (vs 15.4%).
- The big asymmetry is at the level of businesses: just 8.9% of Greek businesses with 10 or more employees were using an AI technology compared to 20% in the EU-27. Greece had the lowest level of enterprise adoption of all the Southern European countries considered.
- Age remains a major divide with AI use being documented at 87.4% among the 17-24 year old cohort and just 34% of those 65 and over.
- Education strongly shapes access to AI’s benefits: Use reached 50.3% among respondents with up to secondary education, 71% among tertiary-education graduates and 89.4% among holders of a master’s degree or doctorate.
- Income inequality is still very much on show: AI usage went from 55.2% in households earning less than €1,000 a month to 85.1% in those earning more than €3,000.
- There’s still a gender gap: In 2026 70.5% of men said they used AI compared to 59.8% of women.
- There’s strong support for regulation: 90.3% of those familiar with AI supported some form of regulation, while only 8.9% favoured completely unregulated use.
- Greece is markedly more anxious about workplace AI than the EU: Only 48% viewed the use of robots and AI at work positively, compared with 62% in the EU. Moreover, 80% expected more jobs to be lost than created, compared with 66% across the EU.
- Privacy remains a major concern: Approximately 71.5% expected AI to have a negative or rather negative impact on privacy protection.
- Greek society is adopting AI faster than the country’s businesses, institutions and productive structures can meaningfully absorb it. Without coordinated policies, AI may reproduce or intensify inequalities that existed before its diffusion.
