India should avoid the global race to build cutting-edge artificial intelligence models and instead focus on using AI to solve practical economic problems, the Economic Survey 2025–26 has argued. At a time when AI adoption is accelerating worldwide, the Survey makes a clear case for a bottom-up, application-led strategy suited to India’s labour market, resource constraints, and strategic interests.
Artificial Intelligence is no longer a future concept. It is already reshaping how firms operate across the world. A McKinsey survey of 1,993 firms showed that by 2025, 88 per cent of organisations were using AI in at least one business function. Among them, 31 per cent were scaling AI across their operations, while 7 per cent had fully integrated it.
AI adoption remains concentrated in high-income countries, which account for 58 per cent of global usage. However, uptake in upper- and lower-middle-income countries is rising steadily. This widening adoption frames the choices before India.
The Economic Survey 2025–26 makes a clear argument: India should not copy the path followed by the United States or Europe. Those economies have pursued frontier AI models backed by massive private capital, large data centres, and the concentration of intellectual property in a few firms. For India, the Survey notes, this model would be costly and inefficient.
Instead, it argues that “India’s comparative advantage in the AI era does not lie in replicating frontier-scale model development”, but in application-led innovation that solves real economic problems.
The Survey addresses widespread fears that AI will destroy jobs. It points to early evidence from the United States showing no “discernible disruption” in the overall labour market. Other studies find that differences in job prospects between occupations with high and low AI exposure are limited. Evidence from Denmark also suggests that most workers have benefited from AI adoption.
This is reassuring for a labour-rich economy like India. However, the Survey warns that this does not justify complacency. It notes that while AI can augment labour in the short term, “productivity gains from augmentation have a ceiling”. Over time, the link between output growth and employment growth may weaken, creating uncertainty about future job creation.
For example, AI can help office workers process documents faster or assist engineers in analysing large datasets. But once these tools are fully embedded, productivity gains may slow, while demand for some roles may decline. The Survey calls this “one of the most considerable looming uncertainties” facing policymakers.
Beyond jobs, the Survey raises a strategic concern. Heavy reliance on foreign AI systems could leave India exposed to geopolitical pressures. It notes that, like semiconductors and critical minerals, AI capabilities will increasingly be used as tools in geostrategic negotiations.
As a result, AI should not be seen merely as a technological upgrade. The Survey states that it has “far-reaching implications for India’s critical infrastructure, labour market, foreign policy and culture”.
India enters the AI era with several strengths. It is among the top global contributors to AI research, has a large pool of technical talent, and is one of the most AI-literate workforces in the world. It also generates vast domestic data across sectors such as health, agriculture, finance, education, and public administration. The Survey notes that this data advantage remains underused.
Bottom-up approach
Globally, two AI development paths are emerging. Western firms have followed a top-down approach centred on frontier models and heavy capital investment. Elsewhere, a bottom-up model has taken shape, with distributed innovation across sectors and strong state coordination.
The Survey argues that “India’s position in this landscape makes a bottom-up approach strategically necessary.”
Practical applications illustrate this logic. In agriculture, AI can help predict weather patterns or identify crop diseases. In healthcare, AI tools can assist doctors in screening X-rays in district hospitals. In manufacturing and logistics, AI can optimise supply chains and reduce waste. These gains do not depend on frontier models but deliver tangible economic value.
Regulation
The Survey argues that India must adopt a risk-based and phased approach. Regulation should evolve alongside deployment, not after it.
It proposes stronger data governance that preserves openness to cross-border flows while ensuring oversight and domestic value retention. Large-scale use of Indian data should remain auditable, with mirrored datasets maintained in India rather than rigid localisation mandates.
The Survey also calls for the creation of an AI Safety Institute to assess risks, conduct testing, and improve transparency. It stresses that public disclosure of safety evaluations must be non-negotiable, as information gaps currently favour large technology firms.
The survey argues that India’s AI strategy should proceed in phases. The first focus should be coordination and experimentation. The next focus is to scale successful applications and introduce proportionate regulation. Over the long term, India must build resilience in computing access and adapt education and skills systems.
