Where commercial incentives are weak but social returns are high, the State must lead AI deployment. Discuss with examples from agriculture and health.

The World Development Report 2026 finds that developing countries gain most by adapting AI to local conditions, not just adopting it [1]. At the report's India launch, World Bank Chief Economist Michael Kremer said the State has a role where commercial incentives are weak but social potential is high [2]. Agriculture and health are the clearest cases.

Why markets under-provide

• Public-good nature: weather forecasts and disease screening benefit everyone, so private firms cannot easily recover their costs.
• Users who cannot pay: smallholders and poor patients have little purchasing power and low digital literacy.
• High fixed costs: compute and datasets are costly. The IndiaAI Mission (₹10,371.92 crore) funds both [3].

Agriculture

• Monsoon forecasting: meteorological agencies and the Agriculture Ministry blended several AI models, including open-access ones. The result predicted the unusual 2025 monsoon, and Ethiopia is now following India's model [2].
• Last-mile access: farmers who have never used a computer get voice advisories by phone. In Gujarat, AgriStack reaches 5 lakh farmers and Sarlaben supports the health of 3 crore cattle [2].
• Baramati: sugarcane yields rose 40% with 30% less water and fertiliser [2]. This is a pilot the State can scale.

Health

• AI tools raised TB detection by 35% [2]. Screening poor patients at that scale is not commercially viable.
• Earlier diagnosis cuts transmission. Markets ignore this kind of spillover benefit, but public budgets should count it.

Limits of State leadership

• Delivery gap: IndiaAI Mission fund utilisation has stayed below 50% since launch [4].
• Trust risks: AI may embed bias in government decisions, erode data privacy and concentrate market power [5].
• Weak foundations: electricity, connectivity and skills must come first [1].

So the State's role is to enable, fund and act as first buyer where markets fail, not to replace private innovation. Faster fund release, open models, bias audits under the Safe & Trusted AI pillar and independent evaluation of pilots can turn successes like Baramati into national programmes. Done well, State-led AI can advance SDG 2 (Zero Hunger) and SDG 3 (Good Health) and make AI's promise inclusive.