AI reshapes South Asia's jobs, creating risks and new opportunities - The Kathmandu Post

TEXT START: Artificial intelligence is beginning to reshape South Asia’s labour markets, slowing hiring in some sectors while creating new opportunities for workers and service exporters, according to a new World Bank report.

The Dissection

The article is a transition-management document. It acknowledges hiring deceleration, reshoring risk, and the exposure of BPO and ICT work, then redirects attention toward AI skills, productivity gains, higher-value exports, and expanding technology expenditure.

Its central move is to confuse growth in output and export revenue with preservation of mass employment. The text identifies the fracture but frames it as an upgrade cycle: routine workers supposedly move upward, suppliers supposedly capture more valuable work, and the region supposedly converts automation into development.

The Core Fallacy

The article treats AI complementarity as a durable employment solution. Under the Discontinuity Thesis, complementarity is not the same as indispensability. If AI makes one skilled worker capable of producing what previously required ten, demand for the service can grow while total hiring still falls.

The reported wage premium for AI skills describes scarcity during the transition, not a permanent human moat. Once those tools, workflows, and validation methods become standardized, the premium itself becomes an incentive for further automation. “AI-complementary” work is often the staging ground for AI replacement.

The article also treats South Asia’s low exposure rate as protection. It is mostly a lag indicator. Agriculture, manual work, and low-skill labor remain less exposed to current cognitive automation because they are outside its immediate reach, not because they are structurally secure. The first major wound lands on the moderately educated young workers concentrated in the region’s export-oriented service pipeline.

Hidden Assumptions

• Global technology spending will translate into proportionate employment and export demand for South Asian workers.
• Foreign buyers will continue outsourcing rather than use AI to reshore exposed functions.
• New, higher-value AI work will scale fast enough to absorb workers displaced from routine services.
• Workers can acquire scarce AI capabilities faster than firms can automate the underlying tasks.
• Human testing, validation, domain knowledge, and solution design will remain labor-intensive rather than become AI-managed processes.
• Productivity gains will be distributed through wages instead of accruing mainly to foreign buyers, local firms, and owners of AI infrastructure.
• South Asian suppliers will retain a cost advantage after AI reduces the value of geographic labor arbitrage.
• Aggregate regional growth will compensate for concentrated sectoral collapse.
• Institutional capacity, education systems, infrastructure, and energy supply can support rapid adoption at scale.

None of these assumptions is established by the article. Most are treated as policy aspirations disguised as forecasts.

Social Function

Classification: partial truth, transition management, and ideological anesthetic.

The partial truth is real: AI can raise productivity, create narrow new specialties, and expand demand for some services during the transition. The anesthetic is the implication that these gains preserve the existing development model. The article turns a distributional rupture into a skills shortage and presents retraining as the bridge across a structural chasm.

Its own evidence contradicts its optimism: exports are still growing while hiring slows. That is the early signature of labor compression—more output, fewer workers, greater leverage for firms, and weaker bargaining power for everyone outside the scarce technical tier.

The Verdict

South Asia is not being rescued by AI; its export model is being repriced. AI can enlarge the region’s revenue pool while shrinking the labor share required to generate it. The winners will be owners of AI-enabled firms, infrastructure, energy, logistics, and scarce verification or domain capabilities. The losers will be the broad middle of young, moderately educated service workers who were supposed to climb through BPO and ICT employment.

Under the hardened framework, P1 is already visible in exposed cognitive services. P2 means firms will not preserve human-only work at scale merely to maintain employment. P3 follows when productivity growth outruns the creation of economically necessary human roles. The article describes the opening phase of that process, then labels the narrowing escape routes “opportunities.”