Applied Frontier

Frequently asked questions · Last updated November 2026

AI, jobs and skills in the developing world

Direct answers, each written to stand alone in under 60 words, to the questions ministries, multilateral agencies and employers ask about jobs, skills and training now that AI is changing what work pays for.

Section 1

The developing world

Which countries does Applied Frontier publish country reports on?
Kenya, India, Bangladesh, Indonesia and Brazil. Each report covers how many jobs the country needs each year, how exposed it is to AI, its biggest untapped AI opportunities, what it costs to create a professional today, and what an AI-economy job would cost to train for. Reports are updated every year.
What does it cost to train someone for an AI-economy job in a developing country?
On Applied Frontier’s benchmark, about $760 to $1,510 in Kenya; $240 to $480 in India; $340 to $680 in Bangladesh; $460 to $930 in Indonesia; $2,440 to $4,870 in Brazil. That is roughly one-eighth to one-sixteenth of the cost of a four-year degree in each country. The benchmark applies public teaching cost per student to a three-to-six-month program and covers training only.
Is AI a threat or an opportunity for Africa?
Less of a threat than in rich countries, because most work is physical or informal; a large opportunity if the country trains the roles AI creates rather than the roles it replaces.
What is a Super Intelligence enablement hub?
A local institution that helps businesses and public agencies adopt and run advanced AI, gives technology providers a local partner, and trains the people who operate the systems. Hubs earn fees for each service, so public or development finance seeds them and commercial revenue sustains them.
Do AI-era jobs need a university degree?
No. The eight job families a developing country can build in three years are all trainable in weeks to months and credentialed by demonstrated work.
How should a multilateral agency fund AI-era workforce programs?
On outcomes, with cost per placed worker as the unit of account: short programs in the job families that are hiring, paid in part on verified placement and retained earnings, delivered by providers that carry part of the risk.
Who is behind Applied Frontier?
A founding team with decades of experience in global workforce development, technology, enterprise AI enablement, virtual labor markets, global development, education policy and emerging markets. The team has built and run training-to-jobs programs and works with governments, multilateral agencies and employers.

Section 2

What is changing

What is Applied Frontier?
Applied Frontier is a research, policy and venture lab that helps countries turn AI into jobs. It identifies the job families the AI economy is creating, runs targeted pilots that train and place people in them, and publishes what it costs to create a job. It works with ministries, multilateral agencies and employers.
Why does global education policy need an overhaul for the AI era?
Because the three things it delivered, portable skills, a shared understanding of what a degree means, and a stable price for those skills, all stopped holding between 2023 and 2025, when AI commoditized the highest-paid cognitive skills.
What is the talent supply chain?
The talent supply chain is the system by which universities certify skills, employers price them and workers carry them across firms and borders. It worked for seventy years because the certification was readable. AI has changed which skills are scarce, which gives countries the chance to rebuild it faster and at lower cost.
Which skills has AI commoditized?
Quantitative analysis, software engineering and high-quality writing, the three skill families that commanded the largest wage premiums in white-collar work until roughly 2024.
Can AI really outperform a McKinsey consultant?
On research, synthesis and deck production, yes. Even a previous-generation model produces work that outperforms a seven-to-ten-year consultant with a Harvard MBA on those tasks, at a marginal cost measured in cents.

Section 3

Jobs, skills and signals

Which jobs will the AI economy need in the next three years?
Jobs that specify problems, verify AI output in a domain, own outcomes, deploy and maintain AI systems, steward proprietary data, persuade humans, and operate in the physical world. Applied Frontier publishes a job-family map on three- and six-year horizons.
How will AI-era jobs be priced?
Unknown, and that is the point. Commoditized skills have collapsed to near zero premium; the premium for judgment, accountability and domain context has not yet been set, and whether it prices globally or locally is an open research question.
If a degree no longer signals skills, what will employers use instead?
Today, the type of work and the type of employer on a résumé act as the proxy. The harder problem is reading an individual’s real contribution through the jargon used to describe it; verified work artifacts and demonstrated tasks are the leading candidates.
What skills should we stop training?
Routine quantitative analysis, first-draft writing, entry-level coding, literature review, slide and document production, rote translation, and memorization-based credentialing. These are not worthless; they are free, so paying to learn them is the error.
Is Capital × Labor = Productivity still valid?
Not in its familiar form. AI is capital that substitutes directly for cognitive labor, so the equation needs a term for AI capital that behaves like labor and a depreciation rate on human skills tied to model capability.

Section 4

The cost of college

How much does the US spend on higher education?
On the order of $700 billion a year across all postsecondary institutions, producing roughly four million degrees a year.
Is higher education training people for jobs that will not exist?
In large part, yes. A student who enrolls in fall 2026 graduates in 2030, and no university can currently name the jobs that student is being prepared for.
How long does it take a graduate to earn $200,000 a year?
For most graduates, never at that level. $200,000 was the exchange rate for a top MBA or senior engineer, and both skill bundles have been commoditized.
What does the solution look like?
At a minimum, producing market-needed skills at a small fraction of today’s cost: months not years, thousands not hundreds of thousands of dollars, credentialed by demonstrated work and revised whenever the models improve.