US vs Iran: AI Job Risk in Two Rival Economies

Same peak score, completely different deployment reality

The shared 8.5/10 peak for clerical workers is the most important number in this comparison - and the most misleading if read in isolation. AI exposure scores measure task susceptibility: whether the core tasks of a role fall within current AI capability. They do not measure employer capital, technology access, economic incentive, or the political environment governing whether those tools can be purchased and deployed. On every one of those dimensions, the US and Iran diverge sharply.

US GDP per capita reached $90,026 in 2025 (World Bank Open Data). Iran's stood at $3,924 in the same year - a 23x gap. Enterprise AI deployment is capital-intensive. Software licences, cloud infrastructure, system integration, retraining, and the organisational restructuring required to replace clerical workflows with AI all carry upfront costs. American employers in financial services, insurance, healthcare administration, and legal services have both the capital and the competitive pressure to move fast. Iranian employers, operating under the world's most extensive unilateral sanctions regime with limited access to the US dollar financial system and blocked from purchasing most Western technology, face an entirely different set of constraints.

Iran's ILO ILOSTAT data (CC BY 4.0, 2024 data year) covers 24,075,550 workers across all major ISCO-08 occupation groups. The US Bureau of Labor Statistics OEWS May 2025 release covers 143,066,500 workers. Together that is 167.1 million workers whose AI exposure can be compared directly on the same scoring framework.

Side-by-side: occupation groups in both countries

The table below compares both countries across all major ISCO-08 occupation categories using ILO ILOSTAT data for Iran (2024 data year) and BLS OEWS May 2025 data for the US, both accessed via ILO ILOSTAT (CC BY 4.0). Iran wage data is not available at the ISCO-1 level in the ILO ILOSTAT series for this data year.

The ISCO-08 framework produces identical AI exposure scores across both countries because task susceptibility is an occupational property, not a national one. What differs radically is the proportion of workers in each group, the wages they earn, the economic environment that determines when tasks get automated, and the geopolitical environment that determines which tools employers can access at all.

Why the average gap is 5.07 vs 3.71 - Iran's workforce structure

The 1.36-point gap between US average AI exposure (5.07/10) and Iran's (3.71/10) is explained almost entirely by where Iran's workers are concentrated. Iran's two largest occupation groups are craft and related trades workers (4,572,400 workers, score 2.5/10) and service and sales workers (4,047,900 workers, score 3.5/10). Together those two groups account for 8.62 million Iranian workers - more than a third of the total workforce - concentrated at the lower end of the AI exposure spectrum.

Iran also has a substantial skilled agricultural workforce of 2,534,900 workers scoring 3.0/10 (ILO ILOSTAT 2024). This reflects Iran's agricultural economy, which employs a meaningfully larger share of the working population than the US (where skilled agriculture accounts for just 889,600 workers). Craft, trades, and agricultural work is fundamentally physical and variable in environment - tasks that resist current AI capabilities and will continue to do so for years even in technologically advanced economies, let alone one where AI infrastructure access is constrained by sanctions.

By contrast, the US workforce is concentrated in knowledge-economy and service-economy roles that score much higher. The US professionals group alone covers 42,964,700 workers - nearly double Iran's entire labour force. US clerical workers at 16,488,600 are nearly 20x Iran's clerical group of 843,800. This tilt toward white-collar, administrative, and professional occupations pushes the US average well above Iran's.

The sanctions dimension: why AI access is not just an economic question

Iran operates under one of the most comprehensive sanctions regimes in the world. US sanctions administered by the Office of Foreign Assets Control (OFAC) prohibit most US technology exports to Iran, including cloud computing services, enterprise software, and AI platform access from US-headquartered providers. The Export Administration Regulations (EAR) additionally restrict hardware exports, including the high-performance GPUs and AI accelerators required for on-premise AI deployment at scale.

This means that even for Iranian employers who have the economic incentive to automate - and some in oil, gas, and financial services do - the primary AI tools available to US employers are either legally inaccessible or practically inaccessible for Iranian counterparts. Leading AI platforms from US providers including enterprise-grade large language model APIs are subject to these restrictions. Iran has responded by developing domestic AI programs under state direction and pursuing technology relationships with countries outside the Western sanctions framework, but domestic frontier AI capability remains limited compared to the US commercial ecosystem.

The practical effect for Iranian workers is that even occupations that score 8.5/10 on task susceptibility face no near-term deployment risk of the kind already visible in US financial services and insurance. Iranian clerical workers are exposed in principle. They are not exposed in practice over any 1-3 year horizon, because their employers lack access to the tools that would actually replace them. This decoupling of task exposure from deployment risk is specific to heavily sanctioned economies and does not appear in the raw scores.

Iran's female labour force participation: the structural invisibility

Iran's female labour force participation rate of 14.01% (World Bank 2025) is one of the world's lowest. For context, the global average is approximately 47% and the US female participation rate is 56.3% (World Bank 2025). This has a direct and important consequence for how AI risk is distributed within Iran: it falls almost entirely on male workers.

Of Iran's 24.1 million tracked workers, the overwhelming majority are male. Clerical occupations in many countries tend to have higher female representation - but in Iran, the extreme skew in labour force participation means the 843,800 Iranian clerical workers at peak AI exposure (8.5/10) are predominantly men. This is the opposite of the US pattern, where women represent a larger share of administrative and clerical roles.

The 14% female participation figure also constrains Iran's aggregate AI exposure score from another angle: many of the higher-scoring professional and knowledge-economy roles that would push the average upward in other countries are underrepresented in Iran's labour market precisely because women participate so infrequently. Iran's workforce structure - dominated by male workers in physical trades, agriculture, and industry - inherently tilts toward lower AI exposure categories at the aggregate level.

Economy context: US vs Iran side by side

The table below puts both countries' economic indicators side by side using World Bank Open Data (CC BY 4.0, 2025) and UNDP Human Development Report 2025 (HDR 2025, 2023 data year, licence CC BY 3.0 IGO).

Iran's 8.3% unemployment rate (World Bank 2025) deserves specific comment. It is more than twice the US rate. High unemployment with low female participation suggests a labour market that is structurally constrained rather than at capacity. Iranian employers do not face the same labour shortage pressure to automate that some US employers cite as a driver of AI investment. When workers are already surplus and cheap, the automation ROI case weakens further - especially given the absence of the leading AI tools that make the business case work in the US market.

Iran's HDI of 0.799 (UNDP HDR 2025, rank 75, 2023 data year) reflects reasonably strong health and education outcomes by global standards. Iran has historically invested in university education and technical training. But a well-educated workforce in an economy cut off from global technology markets and constrained by high unemployment faces limited options for channelling that capability into AI-adjacent productivity growth.

The safest jobs from AI in both countries

At the bottom of the AI exposure spectrum, both countries converge. US elementary occupations score 2.0/10 - 3,846,500 workers earning a median of $37,020 per year (BLS OEWS May 2025). Iran's elementary occupations also score 2.0/10, covering 3,417,300 workers (ILO ILOSTAT 2024). Physical tasks in variable, unstructured environments resist automation in both countries equally.

Iran's craft and trades sector at 4,572,400 workers is particularly notable - it is Iran's single largest occupation group, representing roughly 19% of all tracked employment (ILO ILOSTAT 2024). Iran's industrial and construction economy, including large petrochemical and energy infrastructure, sustains a proportionally large physical-sector workforce. These workers score 2.5/10 on AI exposure. They are also - importantly - the group most insulated from displacement even if AI deployment accelerates, because the physical variability of their work on construction sites, in workshops, and in energy facilities exceeds what current and near-term AI robotic systems can handle cost-effectively.

What this means for workers in both countries

For US clerical workers, the question is no longer theoretical. AI workflow tools are already deployed across US financial services, insurance, healthcare administration, and legal support. The 16,488,600 US clerical workers scoring 8.5/10 are in roles where AI augmentation and - in some cases - replacement is happening in 2026 and will accelerate through 2027 and 2028 in the highest-investment sectors. The risk velocity assessment of "disruption imminent (1-3 years)" reflects observed deployment patterns in firms that have publicly reported AI-driven headcount reduction in back-office functions.

For US professionals (42,964,700 workers, score 6.5/10, median wage $82,032 per year from BLS OEWS May 2025), the picture is more nuanced. AI tools augment rather than replace in most professional workflows for now, but the pace of capability development means roles in legal research, financial analysis, and medical documentation face meaningful near-term changes to task composition. Workers in these roles who adapt their skills to work alongside AI tools are better positioned than those who treat their workflow as static.

For Iranian workers, the exposure scores are a longer-range signal. Iran's 843,800 clerical workers at 8.5/10 are technically in the highest-risk category. But the combination of low wages relative to AI tooling costs, blocked access to leading Western AI platforms, a domestic AI ecosystem that has not yet produced enterprise-grade workflow automation tools at scale, and an 8.3% unemployment rate that reduces the business case for automation all push realistic displacement timelines far beyond the 1-3 year window visible in the US. A 5-10 year horizon is more realistic for Iranian clerical disruption, and that assumes both that geopolitical conditions shift and that domestic AI alternatives develop substantially.

Iran's craft workers (4,572,400 at 2.5/10) and agricultural workers (2,534,900 at 3.0/10) face the lowest risk of any group in either country. Their work is physical, variable, and dependent on local knowledge of conditions that AI systems handle poorly. For these 7.1 million Iranian workers, the AI risk question is not relevant in any near-term planning horizon.