India vs Sri Lanka: Which Workforce Faces Greater AI Risk in 2026?

Two very different scales, one shared vulnerability

India and Sri Lanka are geographically close - separated by 65 kilometres of ocean at the Palk Strait. Their labour markets look nothing alike. India employs 476.6 million people according to ILO ILOSTAT 2025 Labour Force Survey data. Sri Lanka employs 7.8 million according to ILO ILOSTAT 2024, sourced from the Department of Census and Statistics Labour Force Survey. That is a 61-to-1 ratio in workforce size.

What they share is the same peak AI exposure score: 8.5/10 for clerical support workers. Both countries have a category of workers - data entry clerks, schedulers, correspondence handlers, filing staff - whose core tasks are exactly what large language models and AI automation tools now do cheaply and fast. The peak risk is identical. The distribution of that risk across the workforce is not.

The surprising finding from the ILO data: Sri Lanka's average AI exposure score (3.53/10) is higher than India's (3.26/10). This is counterintuitive given that India has the larger, faster-growing economy. The explanation is structural: India's enormous agricultural workforce (161.9 million at 3.0/10) and elementary occupations (103.0 million at 2.0/10) pull its average down. Sri Lanka's workforce is more urbanised relative to its size, with proportionally more workers in professional, technical, and clerical roles.

Side-by-side: all occupation groups compared

The table below shows every ISCO-08 occupation group for both countries. India data is from ILO ILOSTAT 2025. Sri Lanka data is from ILO ILOSTAT 2024, with wages sourced from the 2023 DCS survey cycle. Sri Lanka also reports armed forces occupations separately (41,000 workers, 2.5/10), which India does not report at this aggregation level.

Source: ILO ILOSTAT (CC BY 4.0). India: 2025 Labour Force Survey. Sri Lanka: 2024 LFS from Department of Census and Statistics, wages from 2023 cycle. Armed forces occupations (Sri Lanka: 41K at 2.5/10) not shown above; not separately reported for India at this aggregation.

Why Sri Lanka scores higher on average despite a smaller economy

Sri Lanka's GDP per capita was $5,002 in 2025 (World Bank Open Data). India's was $2,702. By this measure, India is the poorer country. Yet Sri Lanka's AI exposure average is higher. The apparent contradiction resolves when you look at workforce composition rather than just national income.

Sri Lanka's largest occupation group is elementary occupations at 1.96 million workers (25.0% of the workforce) - but its next largest groups are service and sales workers (979K, 12.5%), skilled agricultural workers (1.09M, 13.9%), and craft workers (1.08M, 13.8%). Critically, Sri Lanka has a proportionally larger professional and technical layer than India relative to total workforce size. Sri Lanka's professionals represent 7.5% of employed workers; India's professionals represent 5.8%. That difference, combined with Sri Lanka's 3.6% clerical worker share versus India's 2.3%, pushes Sri Lanka's weighted average up.

The explanation is Sri Lanka's education system. Adult literacy in Sri Lanka is 92.7% versus India's 78.2% (World Bank, 2024 data). Expected years of schooling in Sri Lanka are 13.1 years versus 13.0 years for India (UNDP HDR 2025, 2023 data year) - nearly identical at the top level, but mean years of schooling is 10.8 in Sri Lanka versus 6.9 in India. A more educated population channels more workers into the clerical, professional, and technical roles where AI exposure is highest. That is the structural mechanism driving Sri Lanka's higher average score.

The tea plantation workers: the clearest safe-from-AI story

Sri Lanka's 1.09 million skilled agricultural workers include a significant share of tea estate workers - a role the country has built much of its export identity around. Tea picking in Sri Lanka's hill country (Nuwara Eliya, Kandy, Badulla districts) is done almost entirely by hand. Plucking requires the worker to select two leaves and a bud from each shoot, navigating uneven terrain on steep slopes, adjusting pressure based on leaf maturity that varies plant by plant.

The skilled agricultural group scores 3.0/10 on AI exposure - the same as India's agricultural workers. The AI score here is accurate: AI can help with crop disease detection (image recognition), yield prediction (machine learning on weather and soil data), and logistics scheduling. But the physical plucking itself is not automatable at current technology levels for tea at commercial scale. Robotic tea harvesting has been in prototype stages for over a decade in Japan and China with limited commercial deployment. On the steep, irregular slopes of Sri Lanka's tea country, the economics and the technical challenge are even harder.

This contrasts directly with clerical workers scoring 8.5/10. A tea estate clerk recording picker output, managing payroll, or handling correspondence between the estate and the Colombo tea auction is in the high-risk group. The picker is not. AI disruption in Sri Lanka's tea industry is likely to concentrate on the administrative layer, not the field workers.

Explore the full Sri Lanka data, including the agricultural breakdown, at the Sri Lanka interactive data page.

The 2022 economic crisis: a data reliability note

Any analysis of Sri Lanka's labour market must acknowledge the 2022 sovereign debt default and economic crisis. Sri Lanka defaulted on its external debt in April 2022 - the first such default in the country's history. The crisis brought foreign exchange shortages, fuel rationing, medicine shortfalls, and mass protests that eventually resulted in the resignation of President Gotabaya Rajapaksa.

The ILO ILOSTAT 2024 data used here reflects the Department of Census and Statistics Labour Force Survey conducted as the economy was in early recovery. Sri Lanka received an IMF Extended Fund Facility arrangement in 2023 worth $2.9 billion (IMF, March 2023). GDP growth returned to positive territory in 2023-2024 as the stabilisation programme took effect. But the structural disruption of 2022 - workers leaving formal employment, emigration of skilled professionals, sector contractions - means the 2024 occupation distribution is not a picture of a stable steady state. The composition of clerical and professional workers in particular may undercount actual high-skilled emigration from the crisis period.

The AI exposure scores themselves are not affected by this uncertainty - they reflect occupation task characteristics, not employment levels. But the absolute worker counts for high-exposure groups may be somewhat lower than pre-crisis baselines, meaning the headline risk figures here could understate the exposure of those who remain in formal employment.

The HDI gap and what it means for AI deployment speed

Sri Lanka's HDI of 0.776 (rank 89) places it solidly in the "high human development" band. India's 0.685 (rank 130) puts it in the "medium human development" band. This gap matters for AI adoption beyond just the education channel described above.

HDI captures life expectancy, education, and GNI per capita together. Sri Lanka's life expectancy of 77.5 years (UNDP HDR 2025) versus India's 72.0 years reflects a more developed health system. Its GNI per capita PPP of $12,616 compares to India's $9,047 (UNDP HDR 2025, 2023 data year). On all three dimensions, Sri Lanka has a more developed population base for adopting and being disrupted by technology.

AI deployment requires digital infrastructure - reliable internet, device access, payment systems, and institutional frameworks for managing AI tools. Sri Lanka's higher HDI correlates with faster deployment timelines for the AI tools that affect clerical and professional work. The 283,000 clerical workers in Sri Lanka face a shorter runway to AI displacement than equivalent workers in India, not because Sri Lanka has more advanced AI capabilities (it does not), but because the deployment infrastructure is more mature relative to the size of the economy.

What this means for workers in both countries

For India's clerical workers - 11.1 million people at 8.5/10 AI exposure earning a median of $3,339 per year (ILO ILOSTAT 2025) - the risk is real but the timeline is longer than in richer countries. The cost of AI deployment is the same globally, but the economic return on automating a $3,339/year role is far lower than automating the same role at $40,000/year in the US or UK. This pushes the deployment timeline out to roughly 7 to 12 years for most clerical functions in India, rather than the 3 to 5 years more realistic in high-income markets.

India's 27.9 million professionals face a different calculus. India's IT services and business process outsourcing sector exports to high-income markets - and the AI tools replacing BPO functions are being deployed by clients in those markets, not in India. The timeline for India's offshore IT professionals is closer to 3 to 7 years, driven by client-side adoption rather than domestic economics.

For Sri Lanka's 283,000 clerical workers earning a median of $1,560 per year (ILO ILOSTAT 2024, 2023 wage cycle), the economics of automation look similar to India's on paper - low wages reduce the immediate incentive. But Sri Lanka's post-crisis economic environment includes significant pressure to modernise and streamline government and private sector administration. AI tools that reduce operational costs are politically attractive in a country still managing IMF fiscal conditionality. The deployment pressure may come from a different direction than pure wage economics.

The safest path in both countries is the same as the global pattern: physically demanding work requiring variable-environment adaptation. Sri Lanka's tea pickers, India's construction craft workers, and the agricultural labour forces of both countries sit at 2.0 to 3.0/10 and face minimal direct AI substitution risk in the next decade. See how these patterns compare at the global level in the US AI job risk analysis and the US vs World comparison.

For the India vs Bangladesh comparison covering a different South Asian dynamic - a lower-income country with a large garment manufacturing workforce - see the India vs Bangladesh AI jobs analysis. For the full India vs China comparison, see India vs China: Which Workforce Faces Greater AI Risk?