Applied Frontier

Regions · Last updated November 2026

Where can AI create new jobs in the developing world?

Direct answer

AI can create new jobs in every developing region, but the starting point differs. In Africa the risk is exclusion from the gains. In South Asia it is the outsourcing jobs that AI replaces first. In Southeast Asia it is reaching the informal majority, in Latin America the digital gap, and in the Middle East the distance between Gulf investment and a young workforce short of work. Applied Frontier publishes a playbook for each region and country reports that are updated every year.

Which regions have a playbook?

Five regions, one structure: how exposed the region is, why the next three years matter, which job families to build first, what a ministry should do this year, and where multilateral agencies fit.

Africa

The risk is exclusion from the gains, not displacement.

26%of jobs exposed to AI in low-income countries (IMF)

Africa playbook →

South Asia

The outsourcing and IT jobs that built the export sector are the ones AI replaces first.

15%of jobs strongly complementary to AI (World Bank)

South Asia playbook →

Southeast Asia

Most exposure is in work AI can assist, but nearly 80% of workers are informal.

22.9%of employment in occupations exposed to generative AI (ILO)

Southeast Asia playbook →

Latin America

The gains are real, but about half are blocked by the digital gap.

17Mjobs that could gain from AI but lack digital access (ILO and World Bank)

Latin America playbook →

Middle East

Gulf investment on one side, a young workforce short of work on the other.

26.2%youth unemployment in the Arab States in 2025 (ILO)

Middle East playbook →

Country reports · 2026 edition

Which countries have a country report, and what does an AI-economy job cost to train for in each?

Kenya, India, Bangladesh, Indonesia and Brazil. In each, training for an AI-economy role costs roughly one-eighth to one-sixteenth of a four-year degree, on Applied Frontier’s benchmark. The benchmark covers training only; targeted pilots measure the full cost per placed worker.

Africa · 2026 edition

Kenya

800k–1Mlabor-market entrants a year

Benchmark training cost of an AI-economy role: $760–$1,510

Kenya country report →

South Asia · 2026 edition

India

7.85Mnon-farm jobs needed a year

Benchmark training cost of an AI-economy role: $240–$480

India country report →

South Asia · 2026 edition

Bangladesh

2Mlabor-market entrants a year

Benchmark training cost of an AI-economy role: $340–$680

Bangladesh country report →

Southeast Asia · 2026 edition

Indonesia

1.86Madded to the labor force in a year

Benchmark training cost of an AI-economy role: $460–$930

Indonesia country report →

Latin America · 2026 edition

Brazil

8.1Myoung people outside work and study

Benchmark training cost of an AI-economy role: $2,440–$4,870

Brazil country report →
CountryJobs challengeBenchmark training cost, AI-economy roleCost of a degree
Kenya800k–1M labor-market entrants a year$760–$1,510$12,105
Four-year computing degree
India7.85M non-farm jobs needed a year$240–$480$3,805
Four-year engineering degree, minimum fee
Bangladesh2M labor-market entrants a year$340–$680$5,426
Four-year public-university degree
Indonesia1.86M added to the labor force in a year$460–$930$7,400
Four-year computer science degree
Brazil8.1M young people outside work and study$2,440–$4,870$38,986
Four-year federal-university degree

Benchmark: public teaching cost per student applied to a program of three to six months. Sources and method are in each report. Figures for Kenya, India, Indonesia and Brazil are converted at 2024 average exchange rates and Bangladesh at the September 2026 rate.

Which new job families can a developing country build in three years?

Eight families that sit between a global model and a local outcome. None requires a four-year degree. All can be trained in weeks to months, and most cannot be offshored. Each regional playbook names the four to build first.

3–6 months to train

AI deployment technician

Installs, configures, secures and maintains AI systems inside businesses, clinics, farms and government offices. Local, hands-on, in chronic shortage everywhere.

2–4 months on top of domain experience

Domain verifier

Checks AI output in health, agriculture, finance, law and education before it is acted on. The model is global; the verification is local.

2–3 months to train

Local-language data steward

Collects, labels, governs and licenses data in languages and domains the models lack. The asset is the data, not cheap annotation.

2–3 months to train

Agentic workflow builder for SMEs

Assembles AI agents into working processes for small firms: invoicing, inventory, customer service. Millions of SMEs with no software.

2–6 weeks added to an existing role

AI-augmented frontline worker

Community health worker, extension agent, teacher or clerk operating with an AI copilot. The largest employment pool in any developing economy.

3–6 months to train

Physical-world AI integrator

Sensors, drones, solar and storage, logistics, with AI in the loop. Energy and agriculture are among the largest sectors in most developing economies.

3–6 months to train

AI infrastructure operator

Data-center operations, power, cooling, network. Follows the compute investment arriving in 2025–27.

6–12 months with management experience

Outcome owner

The accountable human for an AI-run process in a bank, ministry or utility. Liability and trust remain human.

Not on the list as a strategy: data annotation, entry-level software development, “prompt engineer” roles and call-center work. Annotation is a task we will place in the best market for it; it is not a national strategy, because it is priced by the buyer and being automated.

How are the regional playbooks and country reports organized?

In three levels. Each region has a playbook. Each country report sits under its region, and each report is updated every year with the same indicators, so that countries can be compared with each other and with their own earlier results.

Five regions

Regional playbook

One page per region: exposure, the three-year window, priority job families, actions for ministries and the role of multilateral agencies.

Five countries

Country report

One permanent page per country, under its region. It always shows the latest edition.

2026 edition

Annual edition

Country data is updated every year with the same indicators, so progress can be compared year on year.

What does each country report cover?

  • How many jobs the country needs each year, from population and labor-force data.
  • How exposed the country is to AI, and where the pressure falls.
  • The biggest untapped AI opportunities, by job family.
  • What it costs to create a professional today, and the benchmark cost of training for an AI-economy role.
  • What the government can do in the next 12 months.

Questions about the regional playbooks and country reports

Which developing regions are most exposed to AI?
Estimates differ by method. The IMF puts exposure at about 26% of jobs in low-income countries and 40% in emerging markets, against 60% in advanced economies. The World Bank’s World Development Report 2026 puts jobs at risk of automation by generative AI at 4.5% in low- and middle-income countries, against 14.2% in high-income countries.
Which regions does Applied Frontier cover?
Africa, South Asia, Southeast Asia, Latin America and the Middle East. Each has a regional playbook covering exposure to AI, the three-year window, priority job families, actions for ministries and the role of multilateral agencies.
Which countries have a country report?
Kenya, India, Bangladesh, Indonesia and Brazil. Each report covers how many jobs the country needs, its exposure to AI, its largest untapped AI opportunities, what it costs to create a professional, and what an AI-economy job would cost to train for.
How often are country reports updated?
Every year. Each country keeps one permanent page that always shows the latest edition, so that progress can be compared year on year.

Sources

  1. IMF, “AI Will Transform the Global Economy. Let’s Make Sure It Benefits Humanity” (January 2024)
  2. World Bank, World Development Report 2026: The Promise of Artificial Intelligence (August 2026)
  3. World Bank, South Asia Development Update: Jobs, AI, and Trade (October 2025), summarized in “Is AI adoption impacting job markets in South Asia?”
  4. ILO report on generative AI and employment in ASEAN (July 2026), as reported by The Star
  5. ILO and World Bank, Buffer or Bottleneck? Employment Exposure to Generative AI and the Digital Divide in Latin America (July 2024)
  6. World Bank, Middle East, North Africa, Afghanistan and Pakistan Economic Update (October 2026), as reported by Khaleej Times
  7. Center for Global Development, “SMS Charges May Be Inhibiting the Use of AI for 1 Billion People” (October 2025)