AI policy must take into account democratic bargain

The share of jobs directly exposed to artificial intelligence is considerably smaller in India than in advanced economies. Yet, its development model is more vulnerable for three reasons.

First, IT services and entry-level cognitive work have been central to India’s post-liberalisation development model and the creation of its middle class. More importantly, these upwardly-mobile professionals constitute India’s consuming class and support employment in other sectors. The consequences of disruption would, therefore, extend beyond those directly employed in IT and professional services.

Second, unlike advanced economies with ageing populations, India has the world’s largest youth population and needs to create productive employment. A technology that allows firms to grow with fewer workers has different implications for a country facing labour shortages and one struggling to absorb young people into work.

Third, the redistributive mechanisms being contemplated in advanced economies, such as universal basic income or public equity stakes in frontier AI companies, are not available to India.

The routine tasks that AI can now perform are the same tasks through which inexperienced workers become experienced. If firms retain senior professionals to review AI-generated work while reducing junior recruitment, they may weaken the pipeline that produces tomorrow’s professionals.

The second channel is exports. A foreign bank or retailer that once outsourced coding, support or document processing to India may perform that work internally with AI and a smaller team. Equally, Indian firms may find new opportunities without translating into employment growth. Our recent YouthPOWER analysis shows that sectoral growth is already diverging from workers’ outcomes. While software exports have continued to grow, employment and real earnings have weakened.

In India, the impact of technology is likely to show up as an increase in lower-quality work. Already, unemployment is substantially higher among young graduates than among young people as a whole. Consequently, a policy that responds only to announced layoffs will miss those who never get hired. At the same time, AI can widen access to scarce expertise, strengthen small firms and improve public services. The task is to shape adoption so that these gains expand productive opportunities.

First, we need a national AI and Work Observatory to track the transition through entry-level recruitment, wages, retention and progression.

Second, we must (re)build pathways into work and protect existing workers. Sectoral apprenticeship arrangements should provide paid, supervised experience using AI on real problems. Generic AI certificates cannot substitute for domain knowledge or professional judgement. Existing workers should have meaningful human review for automated decision-making in the workplace. Portable social protection must accompany this transition.

Third, the transition needs democratic representation. Workers and young people must have a voice in deciding the terms of adoption. A dedicated parliamentary committee should examine AI’s effects on work and hear regularly from young people, workers, small firms and state governments alongside technology companies.

These measures must sit within a wider employment strategy to create demand for labour itself. Interventions like apprenticeships and skilling improve access to existing work, not create new work. Ninety per cent of young workers (85 per cent of all ages) in India earn less than Rs 20,000 a month. India’s depressed wages constrain domestic consumption and limit the ability of firms to grow and hire. It is this that India needs to tackle along with public investments in health, care, education, housing, urban services and labour-intensive enterprise, where unmet needs are substantial, and technology can strengthen workers’ capabilities.

Work gives people competence, independence, social standing and a stake in the future. The democratic bargain rests partly on the expectation that effort can improve one’s circumstances. India must, therefore, judge its AI strategy by whether it expands access to productive work and upward mobility. Higher productivity will be a limited achievement if the routes through which young people can share in it continue to narrow.

The writer is executive director, Future of India Foundation