AI Impact on Employment in India: A Task-Exposure and Scenario Analysis

Abstract

The impact of AI, in particular generative AI, on employment patterns globally is quite different from that in India due to the country's prevalent informal and agricultural sectors. In this paper, a sector-specific task-exposure index for India is constructed based on employment shares in the PLFS and an occupational-exposure model from the international literature and is subjected to three scenarios of displacement. Estimates show that roughly 15 percent of India's total working population are in occupations that are "materially impacted" by AI in the short run, encompassing the services, IT-BPM, finance and clerical occupations, rather than the agriculture and construction sector. It is only under moderate assumptions, however, that net job displacement is relatively small overall (around 1.5 to 6 jobs per 100 workers, across scenarios) and significantly greater for routine-cognitive, white-collar and urban, young workers. The paper recommends building policy that focuses on the support of reskilling, AI-complementary education, and social protection as crucial measures in policy, instead of restricting adoption.