India vs Bangladesh: Which Workforce Faces More AI Risk in 2026?

Two South Asian neighbours, near-identical development levels

India and Bangladesh share a 4,156-kilometre border and a common history. Bangladesh achieved independence in 1971, with India playing a decisive role in supporting the liberation from Pakistan. Fifty years later, both countries sit at an almost identical point on the human development index - a convergence that was not foreordained. Bangladesh was among the world's poorest countries at independence. Its catch-up over five decades, driven substantially by the ready-made garment (RMG) sector and remittance income, has brought it level with India on the composite HDI measure.

That makes this one of the genuinely useful comparisons in the ILO ILOSTAT dataset: two neighbours at the same development level, with different economic structures, facing the same AI disruption pressures. The question is whether those structural differences - India's vast agricultural and IT services workforce versus Bangladesh's garment-heavy manufacturing base - produce meaningfully different AI risk profiles.

The headline answer is: not very different. Both average near 3.2/10 on weighted AI exposure. Both peak at 8.5/10 for clerical workers. Both face distant disruption timelines. But the paths through which AI risk arrives in each economy are structurally distinct, and that matters for which groups of workers should pay attention now.

Side-by-side: all occupation groups compared

The table below uses ILO ILOSTAT (CC BY 4.0) data - 2025 Labour Force Survey for India, Bangladesh Bureau of Statistics (BBS) Labour Force Survey 2024 for Bangladesh. AI exposure scores apply to ISCO-08 occupation groups and are consistent across both countries.

Source: ILO ILOSTAT (CC BY 4.0), India 2025 Labour Force Survey; Bangladesh Bureau of Statistics Labour Force Survey 2024. Wages are median annual values in USD from ILO ILOSTAT earnings dataset.

Why clerks are the highest-risk group in both countries

Clerical support workers score 8.5/10 on AI exposure in both India and Bangladesh - and across every country in the WorldJobsData dataset. That is not a coincidence: clerical work is defined by tasks that AI language models are directly designed to perform. Data entry, document classification, scheduling, correspondence drafting, form processing - these are the core activities of ISCO-08 group 4, and they map almost exactly to the capabilities of current AI tools.

In India, 11.1 million clerical workers earn a median of $3,339 per year (ILO ILOSTAT 2025). In Bangladesh, the same group numbers only 683,800 workers at $2,354 per year (BBS LFS 2024). The wage difference matters for deployment timing: AI tools cost the same to implement regardless of where a worker is located, so the cost-benefit case for automating a $3,339 salary closes faster than for a $2,354 salary. Bangladesh's clerical workforce is both smaller in absolute terms and lower-paid, pushing its automation timeline further out than India's already-distant 10 to 15 year horizon.

India's 27.9 million professionals (6.5/10) are the group where AI disruption is most likely to arrive first, and where the impact is most internationally significant. India's IT services and business process outsourcing (BPO) sector is a major employer of software engineers, analysts, and knowledge workers who earn $5,273 per year on average (ILO ILOSTAT 2025). AI tools that write code, process documents, and handle complex queries are already affecting productivity benchmarks in this sector - not destroying jobs immediately, but changing what one person can produce and compressing headcount growth in offshore services.

Bangladesh's garment sector: an AI story or a robotics story?

The ready-made garment industry is Bangladesh's largest employer and its most consequential economic sector. RMG exports account for more than 80% of Bangladesh's merchandise export earnings, according to the Bangladesh Garment Manufacturers and Exporters Association (BGMEA). The sector employs approximately 4 million workers directly, with the vast majority being women in sewing, cutting, and assembly roles.

These workers fall primarily within the craft and trades workers category (10.2 million total, 2.5/10 AI exposure) and plant and machine operators (6.0 million, 3.0/10 AI exposure). The AI exposure scores are low because sewing, cutting, and physical assembly are not easily automated by language models or digital AI tools. They are physical, dexterous tasks in variable environments.

The honest framing is this: Bangladesh's garment workforce faces a robotics threat, not an AI threat. Industrial sewing robots and automated cutting machines already exist in high-wage garment markets (Japan, Germany). As their costs fall, they become viable in lower-wage contexts. The timeline for robotic displacement in Bangladesh's RMG sector is uncertain but plausibly within 15 to 20 years for the most standardised product lines. AI alone will not threaten these jobs - but robotics combined with falling production costs in competing markets is a structural risk the sector acknowledges.

For workers in these roles, the AI exposure score of 2.5/10 is genuinely reassuring for the medium term. What the score does not capture is robotics risk (scored separately at 4.5/10 for craft workers in the dataset) or the geopolitical and trade policy risks that could shift orders away from Bangladesh regardless of automation.

Agriculture absorbs both workforces - and buffers both AI scores

The single biggest structural feature shaping both countries' AI risk profiles is agriculture. India's ILO ILOSTAT 2025 data shows 161.9 million skilled agricultural workers - 34.0% of the entire workforce - plus 103.0 million elementary workers (21.6%), many of whom work in agricultural or agricultural-adjacent roles. Bangladesh has 29.1 million skilled agricultural workers at 42.2% of its workforce by the BBS LFS 2024 count.

Agricultural and elementary workers score 3.0/10 and 2.0/10 respectively. These groups are not in the direct firing line of AI tools in 2026 or in the foreseeable medium term. Precision-agriculture AI tools (yield optimisation, pest detection, irrigation scheduling) exist and are advancing, but they augment farming operations rather than replacing agricultural labour at scale - particularly at the smallholder farm level that characterises most employment in both India and Bangladesh.

The practical consequence is that both countries' high agricultural employment mechanically pushes down the national weighted-average AI score. That is real information: most workers in India and Bangladesh are genuinely insulated from near-term AI disruption. The risk that exists is concentrated in a numerically small but economically important group - the combined clerical, professional, and managerial class of roughly 65 million people in India and 6 million in Bangladesh.

The HDI parity and what it means for AI adoption speed

Both countries hold an HDI of 0.685 and share rank 130 (UNDP Human Development Report 2025, 2023 data year). That convergence is the most striking feature of this comparison. But the components that compose the HDI differ in informative ways.

Bangladesh outperforms India on life expectancy: 74.7 years versus 72.0 years (UNDP HDR 2025). Bangladesh's female labour force participation rate of 38.65% (World Bank 2025) also exceeds India's 32.42% (World Bank 2025), largely driven by the garment sector's predominantly female workforce. India leads on GNI per capita PPP: $9,047 versus $8,498 (UNDP HDR 2025, 2023 data year), reflecting a larger formal services sector and a more diverse export base.

For AI adoption speed, GNI per capita and the size of the formal digital economy matter more than composite HDI. India's larger IT services industry, wider smartphone penetration, and more developed digital payments infrastructure give it a faster baseline for AI tool adoption among professional workers. Bangladesh is catching up rapidly on digital infrastructure, with mobile financial services reaching rural populations - but the base for enterprise AI deployment is currently smaller.

The wage gap between the two countries also matters: India's GDP per capita of $2,702 (World Bank 2025) versus Bangladesh's $2,597 (World Bank 2025) is less than a 4% difference - genuinely near-peer. Neither country sits at a wage level where AI deployment currently yields large returns for most roles. That is the structural fact that makes both countries' disruption timelines distant despite having clerical and professional groups that are technically highly exposed.

Remittances: an indirect AI exposure channel

Bangladesh is one of the world's largest remittance-receiving economies. The World Bank estimates that remittance inflows to Bangladesh represent approximately 5 to 6% of GDP annually. Millions of Bangladeshi workers are employed abroad - in Gulf states (UAE, Saudi Arabia, Qatar, Kuwait, Oman), Malaysia, and increasingly in other markets. Their earnings support families at home and contribute substantially to foreign exchange reserves.

This creates an indirect AI exposure channel that does not appear in any domestic occupational data. If AI tools reduce demand for Bangladeshi labour in Gulf construction, hospitality, and services markets - or if those economies automate the roles that migrant workers currently fill - the remittance income that supports Bangladeshi households falls. The occupational data for Bangladesh's domestic workforce scores these absent workers' roles at low to moderate AI exposure. But the overseas roles those workers hold are often clerical, service, or semi-skilled positions in richer economies where AI deployment happens faster.

India similarly sends millions of workers abroad - particularly to Gulf states and to high-income English-speaking countries in IT services. The risk to India's overseas IT workers from AI in those destination countries is a faster-moving dynamic than the domestic automation timeline suggests. The India country analysis covers the domestic picture in full; this cross-border channel is an important supplement to that data.

What this means for workers in both countries

For the majority of workers in India and Bangladesh - agricultural workers, elementary occupations, craft trades, and machine operators - the honest message is that AI is not a near-term job threat in 2026. These groups score between 2.0/10 and 3.0/10, and the economics of AI deployment at current wage levels do not justify rapid automation of these roles. The disruption timeline for agricultural and elementary workers is 15 to 20 or more years, subject to substantial uncertainty around how fast robotics costs fall.

For clerical workers in India - 11.1 million people - the timeline is shorter but still distant: 10 to 15 years at current deployment economics. The pressure will show up first not as mass job loss but as slower headcount growth: organisations that previously hired 10 clerks for a growing workload hire 7 or 8 instead, because AI handles the rest. For Bangladesh's 683,800 clerical workers, the same dynamic applies but with a smaller base and even lower wages stretching the timeline further.

For professionals in India - 27.9 million people, many in IT services and BPO - the timeline is 5 to 10 years for meaningful disruption in specific functions. AI coding tools, document automation, and knowledge management systems are already live in global technology firms. The offshore services market that employs many Indian professionals is one of the first places where AI-driven productivity gains are compressing headcount. Workers in these roles should treat upskilling in AI tool use as a near-term professional priority, not a future concern.

Explore India's full occupation breakdown at the India country data page and Bangladesh's at the Bangladesh country data page. Both are updated with ILO ILOSTAT data and include economy indicators from World Bank Open Data and UNDP HDR 2025.