Janmejaya Sinha, BCG’s chairman in India, argues that emerging economies should use generative AI to augment workers’ capabilities rather than prioritize automating their jobs. He builds on World Bank president Ajay Banga’s observation that generative AI will affect emerging markets differently from advanced economies, and asks whether the right choices could help narrow the gap between them.

Generative AI requires advanced chips, substantial computing capacity, reliable electricity, data centres, data, large language models and increasingly sophisticated foundation models. Sinha says emerging economies start at a disadvantage in nearly all these areas. US AI investment runs into the trillions of dollars, exceeding the GDP of many emerging economies; advanced chips are scarce and expensive, and training and operating frontier models demands substantial capital and energy. Leading models come from a relatively small group of companies concentrated largely in the US, while China is also investing heavily.

Sinha argues that emerging markets may gain more by using AI to access expertise at low marginal cost than by trying to produce AI. A dairy farmer who rarely sees a veterinarian, a nurse far from a specialist physician, or a small business owner without an accountant, lawyer or marketing adviser could receive assistance in their own language. The article says fewer jobs in emerging economies are directly exposed to generative AI automation: between 4.5% and 11%. In India, more than 500 million people depend on agriculture. In advanced economies, around 30% of jobs are in clerical, administrative, professional and highly digitised occupations that are more exposed, according to the ILO.

AI could help farmers identify animal disease, adjust feeding and breeding decisions, spot crop pests, and assess irrigation, fertiliser or weather risks without replacing their physical work. In health care, AI-supported nurses, community health workers and caregivers could improve triage, documentation, translation, medication checks and monitoring, allowing less highly trained workers to perform some tasks more reliably and productively. Sinha contrasts this opportunity with ageing populations in Japan, South Korea, China, France and Germany, where health systems face workforce shortages and budget deficits; he says worker displacement could further strain consumption and labour-based tax revenues.

Emerging economies can observe how adoption unfolds in developed economies, where higher wages, stronger digital infrastructure and more AI-exposed jobs may bring difficult questions about automation, job redesign, reskilling, inequality and sharing productivity gains sooner. Sinha says emerging economies should focus on sectors employing many people, local-language applications, voice-first interfaces for workers with limited digital literacy, training, and rules distinguishing decision support from decisions reserved for qualified professionals. He says policy, infrastructure, education, business models and regulation will shape whether generative AI spreads access to expertise and helps narrow the gap with advanced economies; there is no guarantee that it will.

Janmejaya Sinha, BCG’s chairman in India, argues that emerging economies should use generative AI to augment workers’ capabilities rather than prioritize automating their jobs. He builds on World Bank president Ajay Banga’s observation that generative AI will affect emerging markets differently from advanced economies, and asks whether the right choices could help narrow the gap between them.

Generative AI requires advanced chips, substantial computing capacity, reliable electricity, data centres, data, large language models and increasingly sophisticated foundation models. Sinha says emerging economies start at a disadvantage in nearly all these areas. US AI investment runs into the trillions of dollars, exceeding the GDP of many emerging economies; advanced chips are scarce and expensive, and training and operating frontier models demands substantial capital and energy. Leading models come from a relatively small group of companies concentrated largely in the US, while China is also investing heavily.

Sinha argues that emerging markets may gain more by using AI to access expertise at low marginal cost than by trying to produce AI. A dairy farmer who rarely sees a veterinarian, a nurse far from a specialist physician, or a small business owner without an accountant, lawyer or marketing adviser could receive assistance in their own language. The article says fewer jobs in emerging economies are directly exposed to generative AI automation: between 4.5% and 11%. In India, more than 500 million people depend on agriculture. In advanced economies, around 30% of jobs are in clerical, administrative, professional and highly digitised occupations that are more exposed, according to the ILO.

AI could help farmers identify animal disease, adjust feeding and breeding decisions, spot crop pests, and assess irrigation, fertiliser or weather risks without replacing their physical work. In health care, AI-supported nurses, community health workers and caregivers could improve triage, documentation, translation, medication checks and monitoring, allowing less highly trained workers to perform some tasks more reliably and productively. Sinha contrasts this opportunity with ageing populations in Japan, South Korea, China, France and Germany, where health systems face workforce shortages and budget deficits; he says worker displacement could further strain consumption and labour-based tax revenues.

Emerging economies can observe how adoption unfolds in developed economies, where higher wages, stronger digital infrastructure and more AI-exposed jobs may bring difficult questions about automation, job redesign, reskilling, inequality and sharing productivity gains sooner. Sinha says emerging economies should focus on sectors employing many people, local-language applications, voice-first interfaces for workers with limited digital literacy, training, and rules distinguishing decision support from decisions reserved for qualified professionals. He says policy, infrastructure, education, business models and regulation will shape whether generative AI spreads access to expertise and helps narrow the gap with advanced economies; there is no guarantee that it will.