Turkey vs Greece: Which Faces More AI Job Risk?

Same peak, different average - how two Aegean neighbours diverge

Turkey and Greece share a long border, a turbulent bilateral relationship, and - as the WorldJobsData dataset shows - an identical peak AI exposure score. Both countries record 9.0/10 for general and keyboard clerks, placing them among the most exposed occupation groups in the global dataset (Eurostat lfsa_egai2d via ILO ILOSTAT 2025, CC BY 4.0). This is not a coincidence. Both datasets use the same Eurostat methodology for classifying occupations under ISCO-08, and the task profile of general and keyboard clerks - data entry, secretarial work, office administration - is uniformly high-risk regardless of which side of the Aegean a worker sits on.

What differs is the weighted average across all occupations. Greece scores 4.88/10 overall; Turkey scores 4.11/10. That 0.77-point gap is structurally driven. Greece's economy is approximately 60% services by employment share (Eurostat national accounts). Tourism alone accounts for roughly 20% of Greek GDP (ELSTAT, 2024). Shipping, financial services, and professional services add further service-sector weight. Turkey's economy is more mixed: manufacturing accounts for around 20% of GDP (TurkStat, 2024), and agriculture - at 10.46% of tracked workers in the ILO ILOSTAT data - remains a larger share of the workforce than in most comparable economies. These physical occupations score between 1.5/10 and 3.5/10, pulling Turkey's weighted average down substantially.

Side-by-side comparison: major occupation groups

The table below compares both countries across the major ISCO-08 occupation categories using Eurostat lfsa_egai2d data via ILO ILOSTAT 2025 (CC BY 4.0). Employment figures are in thousands.

The scores are strikingly consistent across both countries because the Eurostat data uses the same ISCO-08 classification system and the same AI exposure methodology. Where Turkey and Greece differ is not in which occupations are at risk but in how many workers fall into each risk bucket - and Turkey's sheer workforce size (32.4M vs 4.3M) means its absolute numbers are much larger even when the percentage share of high-risk workers is lower.

The 9.0/10 clerical score - why both countries share the peak

General and keyboard clerks in Turkey total approximately 599,500 workers (1.85% of the workforce), per Eurostat lfsa_egai2d via ILO ILOSTAT 2025. In Greece, this sub-group covers 314,400 workers (7.27% of the workforce). The Greek share is nearly four times larger in proportional terms, which directly contributes to Greece's higher weighted average. A country where 7.27% of workers are in the single most AI-exposed occupation category is going to have a materially higher overall average than one where that group represents 1.85%.

The 9.0/10 score for this group reflects the nature of the tasks involved: data entry, keyboard operation, office correspondence, document management, and secretarial functions. As of 2026, large language models can draft, edit, classify, summarise, and route this type of information at a level that meets or exceeds typical human output speed for routine cases. The displacement pressure on this group is real and already visible in hiring data: Eurostat LFS data for both Turkey and Greece shows clerical employment flat or declining relative to overall workforce growth since 2022, even as total employment in both countries has risen.

Both Turkey's ICT professionals (189,000 workers) and Greece's (59,900 workers) score 8.5/10 - a figure that would have seemed extreme five years ago but reflects the degree to which AI code generation, automated testing, and AI-assisted system design tools have shifted the task profile for software and IT roles. In both countries, senior engineering and architecture work remains substantially human, but junior coding, first-line IT support, and routine system monitoring have all seen AI tools absorb a meaningful share of tasks in 2025 and 2026.

Why Greece's average is higher - the service economy gap

The 0.77-point gap in weighted average AI exposure between Greece (4.88/10) and Turkey (4.11/10) maps directly onto the structural composition of each workforce. Greece's largest single occupation group is sales workers at 13.03% of the workforce (563,600 workers), scoring 5.0/10. Turkey's largest single group is market-oriented skilled agricultural workers at 10.46% (3,389,800 workers), scoring 3.5/10. That single difference - largest group scoring 5.0 versus largest group scoring 3.5 - moves the needle substantially on the weighted average.

Greece's service orientation goes beyond sales. Teaching professionals represent 6.79% of the Greek workforce (293,700 workers, 6.5/10 exposure). Personal service workers - a category that captures tourism, hospitality, and personal grooming - account for 6.39% (276,400 workers, 2.5/10). The tourism industry's hospitality roles are low-AI-exposure because they are physical and interpersonal, which acts as a partial counterweight. But the back-office, administrative, financial, and professional services that support Greek tourism and shipping industries are concentrated in the 7.5/10 to 9.0/10 range, and those sectors employ a large share of Greek knowledge workers.

Turkey's manufacturing workforce adds another buffer. Metal, machinery and related trades workers (1,163,800 workers, 3.0/10) and stationary plant and machine operators (1,039,900 workers, 3.5/10) together represent 6.8% of the Turkish workforce. Greece has far fewer workers in these categories - 95,100 metal trades workers (2.20% of workforce) and 63,900 plant operators (1.48%). This is not because Greece lacks industrial workers; it reflects a smaller absolute workforce and a structural shift toward services that accelerated through the post-2010 debt crisis years, when manufacturing employment contracted sharply.

Turkey's agricultural and manufacturing buffer

Turkey's 3.39 million market-oriented skilled agricultural workers represent 10.46% of the tracked workforce (Eurostat lfsa_egai2d via ILO ILOSTAT 2025, CC BY 4.0). At 3.5/10 AI exposure, this group is a significant low-risk anchor. AI tools do assist Turkish precision agriculture - spraying drones, soil monitoring, irrigation scheduling - but the field labour itself remains physical and resistant to full AI displacement at current technology and cost levels. This pattern mirrors what the dataset shows globally: agricultural employment, particularly in middle-income economies, pulls weighted averages down and provides a near-term buffer against the headline clerical displacement story.

The 1,163,800 metal and machinery trades workers (3.59% of workforce, 3.0/10) and 1,039,900 stationary plant and machine operators (3.21%, 3.5/10) contribute further. Turkey's manufacturing sector - automotive, textiles, chemicals, electronics - is robotics-heavy but not yet AI-heavy in the sense that matters for this dataset. Welding robots and CNC automation have already transformed these roles, but the AI dimension (software-driven decision-making, language-model-based job displacement) is a lower threat to plant operators than to clerks. The dataset scores these groups accordingly.

Turkey's large driver workforce is also worth noting: 1,817,400 drivers and mobile plant operators (5.61%, 2.5/10). Autonomous vehicle technology is advancing, but Turkey's complex urban driving conditions, road infrastructure, and regulatory environment mean the displacement timeline for this group is medium-to-long term - 5 years or more for meaningful headcount impact, per the WorldJobsData risk velocity rating, which places Turkey at "disruption imminent (1-3 years)" at the aggregate level but with significant variation by occupation.

The safest jobs in Turkey and Greece

Both countries share the same lowest-scoring occupation: armed forces occupations, other ranks, at 1.5/10. This covers 3,600 tracked workers in Turkey and 3,800 in Greece (Eurostat lfsa_egai2d via ILO ILOSTAT 2025). The tasks in this group - physical field duties, direct supervision of personnel, tactical operations - are by definition outside AI's current capabilities and are legally restricted to human actors. The 1.5/10 score is the floor in both datasets.

Below the armed forces category, both countries share the same low-scoring physical occupations at 1.5/10: cleaners and helpers (Turkey: 1,232,200 workers; Greece: 107,400 workers), agricultural and forestry labourers (Turkey: 1,002,600; Greece: 35,200), food preparation assistants (Turkey: 535,700; Greece: 44,300), street and related sales and service workers (Turkey only: 64,100), and refuse workers (Turkey: 370,700; Greece: 50,300). All of these groups share a common characteristic: tasks that are physical, variable, and performed in real-world environments where AI cannot yet match the cost-effectiveness of human labour.

Building and related trades workers score 2.0/10 in both Turkey (1,151,700 workers) and Greece (145,300 workers). This group - carpenters, plumbers, electricians, general construction workers - faces minimal AI exposure but meaningful robotics exposure (4.0/10). Prefab construction and building information modelling (BIM) are advancing, but on-site trades work in both countries remains predominantly human-executed. The physical judgment requirements of construction in variable conditions, particularly in seismically active regions like the Aegean, make robotic substitution substantially harder than in controlled factory settings.

Economy context: what the GDP and HDI gap means for AI disruption

Turkey's GDP per capita is $18,599 USD (World Bank, 2025 data year), and Greece's is $26,948 USD (World Bank, 2025 data year). Greece's higher income level means AI adoption tends to proceed faster - capital is more available, technology infrastructure is more developed within the EU framework, and labour costs are higher, which improves the cost-benefit case for automation. Turkey's lower GDP per capita means the marginal cost of human clerical labour is lower relative to AI infrastructure investment, which may slow the practical deployment timeline even where the AI exposure score is high.

Greece has an HDI of 0.908 (rank 34 globally, UNDP Human Development Report 2025, 2023 data year). Turkey's HDI is not available in the UNDP 2025 dataset as published, but World Bank life expectancy data (77.42 years, 2024) and adult literacy rate (97.26%, 2021) provide comparable indicators. Greece's stronger HDI score reflects better universal healthcare, higher expected years of schooling (20.85 years vs Turkey's estimated range), and higher mean years of schooling. A displaced Greek clerical worker has better access to retraining programmes within the EU framework - including European Social Fund and Just Transition Fund support - than a comparable Turkish worker, who would rely on ISKUR (Turkiye Employment Agency) schemes with more limited funding.

Both countries have unemployment rates at 8.52% (Turkey, World Bank 2025) and 8.54% (Greece, World Bank 2025) - remarkably close. Greece's recovery from its 2012-2015 peak unemployment above 27% has been substantial; the current figure reflects a genuine labour market recovery, not hidden discouraged workers. Turkey's 8.52% figure includes significant seasonal agricultural employment fluctuation and some underemployment in the informal economy (informal employment rate: 27.71%, 2024). Gini inequality stands at 43.7 for Turkey (World Bank, 2023) versus 33.4 for Greece (World Bank, 2023). The higher Turkish inequality means AI-driven displacement would hit lower-income clerical and service workers harder in relative terms than in Greece, where the income distribution is less extreme.

What this means for workers in both countries

For Turkish and Greek workers in clerical roles - particularly the 599,500 Turkish and 314,400 Greek general and keyboard clerks sitting at 9.0/10 - the question is no longer whether AI can replicate their work. The current generation of large language models can already handle the core tasks of data entry, document drafting, routing, and basic correspondence at a speed and cost that makes automation commercially viable for larger employers. The disruption timeline is medium-term: WorldJobsData rates Turkey's risk velocity at 10.0 ("disruption imminent, 1-3 years") and Greece's at 9.5 (also "disruption imminent"), reflecting that both economies have the technology infrastructure and employer sophistication to begin restructuring clerical headcount in the near term.

The practical difference between Turkey and Greece is in what happens after displacement. Greek workers exist within the EU social protection framework and have access to Active Labour Market Programmes (ALMPs) funded in part through European Social Fund transfers. Retraining pathways, particularly in digital skills and healthcare, are more structured and better-funded in Greece than in Turkey. Turkey's ISKUR employment agency runs retraining programmes, but funding per displaced worker is lower and coverage is more uneven across regions. A displaced clerical worker in Ankara has more limited formal retraining support than a comparable worker in Athens, even though their AI exposure score is identical.

Both countries are experiencing significant population dynamics that interact with AI disruption. Greece has an aging workforce - the World Bank reports life expectancy of 81.84 years (2024) and the UNDP HDR 2025 records expected years of schooling at 20.85. Greece's demographic profile (WorldJobsData rates it as "labor shortage solver" - automation fills critical gaps as the working-age population shrinks) means AI adoption in clerical roles may partly absorb headcount reductions through attrition rather than direct displacement. Turkey, with a younger population structure and life expectancy of 77.42 years, faces a different challenge: a growing workforce entering clerical roles that are becoming automated, requiring job creation in less-exposed sectors to absorb new entrants.