China vs Vietnam: AI Job Risk Compared in 2026

Two communist-market economies, very different AI profiles

China and Vietnam share more than geography. Both are one-party states that adopted market-oriented reforms while keeping communist political structures - China through the reforms of the 1980s, Vietnam through the Doi Moi reforms of 1986. Both have grown rapidly by integrating into global manufacturing supply chains. Both have clerical workers who score 8.5/10 on AI exposure from ILO ILOSTAT data - China from the 2025 National Bureau of Statistics release, Vietnam from the 2024 Ministry of Labour, Invalids and Social Affairs survey.

But the structural difference between their labour markets is significant. China in 2026 has a much larger professional and service-sector workforce - the result of a 40-year economic transition that moved workers up the value chain into finance, technology, and business services. Vietnam is roughly 15 to 20 years behind on that transition. Its workforce is still heavily concentrated in manufacturing, agriculture, and elementary work - the occupations where AI does the least direct damage in 2026.

The result is a 1.21-point average AI exposure gap (4.48 versus 3.27), even though the top of the risk distribution looks identical in both countries.

Side-by-side: all occupation groups compared

The table below shows every ISCO-08 occupation group available for both countries, using ILO ILOSTAT data - China 2025, Vietnam 2024. Vietnam wage data is available from ILO ILOSTAT at occupation-group level. China has no ILO wage data at occupation-group level for 2025.

Source: ILO ILOSTAT (CC BY 4.0), China 2025 NBS release and Vietnam 2024 MOLISA Labour Force Survey. China wage data not available in ILO dataset at occupation-group level. Dash (-) indicates group not separately reported in ILO data for that country.

Why China's average is so much higher

The single biggest driver of China's higher average AI exposure is the size of its professional class. 81.8 million Chinese professionals score 6.5/10 on AI exposure - this group represents 22.6 percent of China's entire workforce, according to ILO ILOSTAT 2025 data. These are engineers, software developers, financial analysts, accountants, legal professionals, and healthcare specialists. They work in roles where AI tools - code generation, document drafting, data analysis, and complex decision support - are already demonstrably capable.

Vietnam's professionals are 3.9 million workers, or 7.2 percent of the workforce. The absolute gap is enormous: China has more than 20 times the number of professionals. Even accounting for China's larger total workforce, the professional share is triple that of Vietnam. This structural difference in how urbanised and service-oriented each economy is - China's GDP per capita was $13,862 in 2025 versus Vietnam's $5,066 (World Bank Open Data) - directly translates into who sits in which occupation tier.

China's 33.6 million clerical workers at 8.5/10 are also substantially larger than Vietnam's 1.3 million. In absolute terms, China has 26 times more workers in the highest-risk occupation group. These are the workers facing the most direct AI disruption timeline - data entry, scheduling, correspondence, and document management tasks that current large language models can handle with high reliability.

The China+1 supply chain shift and what it means for AI risk

From approximately 2018 onward, a meaningful share of global manufacturing shifted from China to Vietnam under the "China+1" strategy - companies diversifying away from single-country supply chain dependence. Apple moved significant iPhone assembly capacity to Vietnam. Samsung built major semiconductor and handset plants there. Nike and Adidas shifted garment production. Vietnam became the world's second-largest exporter of electronics and the third-largest exporter of garments, according to UN Comtrade data.

The workforce consequences show directly in ILO ILOSTAT data. Vietnam's craft and trades workers - the assembly line workers, garment stitchers, electronics assemblers - number 7.2 million at 2.5/10 AI exposure. Plant and machine operators add another 7.6 million at 3.0/10. Together that is 27.9 percent of Vietnam's workforce in occupations that score at or below 3.0/10 - work that is physically located in factories and not susceptible to the kind of AI substitution that threatens office-based roles.

China kept the higher-value work: R&D, software development, financial services, supply chain management, and business process services. Those roles are heavily represented in China's 81.8 million professionals and 33.6 million clerical workers - the occupation groups that account for most of the average AI exposure gap between the two countries. The supply chain shift, in a structural sense, transferred some low-AI-risk work from China to Vietnam while China moved up the value chain into higher-AI-risk service roles.

Explore the full breakdown for both countries: China workforce data and Vietnam workforce data in the interactive tool.

Vietnam's elementary worker base: the clearest low-risk anchor

Vietnam's largest single occupation group is elementary workers - 12.9 million workers at 2.0/10 AI exposure, the lowest score in the ISCO-08 classification. These are manual labourers, agricultural workers, cleaners, and delivery workers. They account for 24.0 percent of Vietnam's entire workforce, per ILO ILOSTAT 2024 data.

Elementary work is where AI does the least damage in 2026. The tasks involved - physical presence, manual dexterity in variable environments, low-complexity but context-dependent actions - are areas where current AI systems have no direct substitution capability. The risk for these workers comes from robotics (scored separately at 5.5/10 for this group in the Vietnam data) and economic displacement pressure, not from AI directly replacing their output.

China's ILO ILOSTAT data does not separately report elementary workers or skilled agricultural workers - these groups are either aggregated or not captured in the national survey at the ISCO-08 level. The occupations China does report - craft workers (93.6M), plant operators (50.0M), service and sales (72.6M) - are all in the 2.5 to 3.5/10 range, but the professional class at 81.8M creates a much heavier weight on the high-risk end of the distribution.

See the full China AI job risk analysis and the full Vietnam AI job risk analysis for the complete country-level breakdowns.

Economy context: China and Vietnam side by side

Economic development level is the most reliable predictor of how fast AI disruption moves through a labour market. Higher GDP per capita means higher wages, which makes automation more economically attractive to employers. Higher HDI means better digital infrastructure, higher digital literacy, and stronger institutional capacity to deploy AI tools at scale.

Vietnam's 1.52 percent unemployment rate (World Bank, 2025) is strikingly lower than China's 4.62 percent. This reflects a near-full-employment manufacturing economy where labour demand from electronics and garment factories has absorbed most of the available workforce. It also means Vietnam has less slack to absorb displaced workers if AI begins affecting its smaller clerical and professional populations.

What this means for workers in both countries

For China's workers, the disruption timeline is already compressing. China's GDP per capita of $13,862 means employers in finance, technology, insurance, and professional services have the capital to invest in AI tools - and the wage levels to make those investments return quickly. The 33.6 million clerical workers and 81.8 million professionals face real AI exposure within a 3 to 7 year window for meaningful job transformation. This does not necessarily mean job losses at the same scale - the ILO and OECD consistently distinguish between task displacement and occupation elimination - but the pressure on these roles is genuine and accelerating.

The safest Chinese workers in 2026 are those in craft and trades roles - 93.6 million workers at 2.5/10. China's construction, manufacturing, and skilled trades sectors employ the largest single occupation group in the ILO dataset for China, and these workers are insulated from direct AI substitution. Their risk comes primarily from robotics in manufacturing environments, not from language model tools.

For Vietnam's workers, the short-term AI risk picture is more favourable. Vietnam's largest groups - elementary workers (12.9 million at 2.0/10), service and sales workers (10.3 million at 3.5/10), skilled agricultural workers (8.2 million at 3.0/10), and plant operators (7.6 million at 3.0/10) - are all in the lower half of the AI exposure scale. The 1.3 million clerical workers who score 8.5/10 are a small enough share of the workforce (2.4 percent) that disruption in this group does not dramatically shift the national average.

Vietnam's medium-term risk, however, is different from the 2026 snapshot. As Vietnam's economy continues to move up the value chain - as it already is in software, fintech, and higher-value manufacturing - its professional class will grow. If Vietnam follows China's economic trajectory, its AI exposure average will follow China's path upward over the next 10 to 15 years. The current 3.27/10 average reflects where Vietnam's economy is today, not where it will be when the next generation of workers moves into white-collar roles.

See how both countries compare to the global picture in the US analysis (155.5 million workers), the India vs China comparison, and the China vs Philippines comparison.