Key Details
The ILO brief, AI and Developing Economies: Promoting Decent Work in a Time of Rapid Technological Change, distinguishes potential exposure to AI, its actual use at work and observed changes in employment. That distinction is essential to reading the figures below.
Jobs potentially affected: About 25% of global employment involves tasks that generative AI could automate or assist. The estimated share is 11% in low-income countries; exposure does not mean those jobs will disappear.
Unequal potential gains: One model projects that, under a high-productivity scenario, AI could raise output over ten years by about 5% in advanced economies, compared with 2.7% in low-income countries. These are projections, not measured gains.
Different exposure for women: Globally, 28% of women’s employment is in potentially exposed occupations, compared with 21.3% of men’s. The gap is much smaller in lower-income countries.
Early hiring signal: A US study cited by the ILO found a 6% employment decline from 2022 to 2025 among workers aged 22–25 in highly AI-exposed occupations, largely associated with reduced hiring. It is not an estimate for India.
Work Behind AI: Data annotation, content moderation and AI-related business process outsourcing create income opportunities, but working conditions vary. Among workers surveyed in 2022–23 on microtask platforms and in AI-related outsourcing firms, 50% reported access to paid holidays and 65% to sick leave. These are findings for the surveyed workers, not India-wide estimates.
The Risk Is Disruption Without a Productivity Dividend
The ILO finds modest, gradual effects on aggregate employment so far. AI is more visibly changing particular tasks and how work is organised than eliminating whole occupations. Yet even where a tool improves an individual task, that gain may not spread across a firm or economy.
Developing economies face a particular mismatch: many have lower overall exposure to automation, but also less reliable connectivity, computing capacity, relevant data and training. Workers in vulnerable digital tasks could therefore encounter AI-driven competition before smaller firms and public services have the capacity to use AI productively. The brief calls this “disruption without dividend.”
Entry-Level Work and Job Quality Need Closer Attention
A stable total number of jobs can conceal changes in who gets hired. The US evidence on younger workers, alongside early findings from online freelance markets, points to pressure on some entry-level writing, coding and customer-service tasks. Those findings are still developing and cannot be assumed to describe India’s labour market. They do, however, identify a question worth monitoring in India’s digitally traded services.
The brief also draws on evidence of algorithmic management, including research on platform work in India. Systems that assign tasks, track performance or influence pay can improve coordination, but may leave workers with little information about decisions affecting them or a way to challenge them.
AI creates another kind of work that is easy to overlook. People label data and moderate content to help build and operate AI systems, often through platforms or outsourcing firms. These jobs can provide income, but repetitive tasks, exposure to harmful material, limited benefits and weak career progression affect whether they amount to decent work. The ILO calls for clear terms, timely payment, mental-health support and ways to challenge arbitrary rejection of completed tasks.
Locally Useful AI Requires More Than Access to a Tool
The ILO sees opportunities for AI that assists workers in agriculture, health, education and public administration. It argues that affordable tools adapted to local languages and needs may matter more to many developing economies than building frontier models. Smaller enterprises will still need practical support—skills, finance, connectivity and access to computing resources—to put such tools to use.
Policy Relevance
India’s AI policy has two connected employment tests. The first is whether smaller firms and workers can use AI to do more valuable work, rather than leaving gains concentrated among well-resourced businesses. The second is whether people whose work is organised or produced through AI retain fair pay, autonomy, protection and a means of appeal.
Neither potential-exposure estimates nor overseas hiring studies can answer those questions for India. Evidence on actual workplace adoption, entry-level hiring, wages, task changes and conditions in platform and AI-support workwould show where assistance is creating opportunity and where safeguards or training are needed.
Follow the Full Policy Brief Here: AI and Developing Economies: Promoting Decent Work in a Time of Rapid Technological Change