Can India S Ai Promise Reach Everyone

Across India, conversations around Artificial Intelligence (AI) are steadily moving beyond innovation labs and corporate boardrooms into the domain of governance. The larger question is no longer whether governments will adopt AI, but whether these systems can genuinely improve public services in ways that are accessible, accountable, and inclusive. It is within the broader shift that Madhya Pradesh recently announced its phased AI Mission across government departments, becoming significant. More than a technological upgrade, the initiative reflects an attempt to explore how predictive systems could support decision-making, strengthen public delivery mechanisms, and respond more effectively to citizen needs.The mission envisions AI not as a back-office efficiency tool but as a system that anticipates crop failures, flags health risks before they cascade, and helps MSMEs access credit they have historically been denied. If that promise is kept, it would represent something genuinely significant: the state using predictive intelligence to serve its citizens.India’s ability to move toward AI-led governance is also being shaped by a digital ecosystem that has expanded rapidly over the last decade. The Digital Public Infrastructure stack with Aadhaar, UPI, ONDC, and the Account Aggregator framework is not merely an engineering achievement; it is a redistributive architecture that has extended formal financial identity to millions of previously excluded citizens. The National AI Mission builds on this base: as of late 2025, over 38,000 GPUs are available at subsidised rates, and AIKosh hosts more than 1,500 public datasets. This is, deliberately, not a market-led model, and that distinction matters because it positions digital infrastructure as a public utility rather than a purely commercial ecosystem.The first generation of AI in Indian governance i.e., chatbots, automated grievance portals, and form-processing bots, was useful but modest. What is now being imagined is categorically different: Decision intelligence. AI systems can help officials simulate the effects of a price support scheme before it is announced, or identify districts most vulnerable to a monsoon shortfall weeks in advance. In agriculture, satellite-based predictive analytics could reduce the volatility that devastates smallholder incomes. In public health, integrating data from pharmacies, hospitals, and environmental sensors can enable pre-emptive surveillance rather than reactive crisis management; in education, finance, and logistics, AI similarly drives personalised learning, smarter credit assessment, and supply-chain optimisation.With over 65% of India’s population under 35, the urgency is demographic as much as administrative. Private-sector decision-making already runs on real-time analytics; public policy still relies on lagged surveys. Closing that gap is not a luxury for a country whose young people need productive employment and functioning health systems in this decade, not the next.None of this is without risk, and the risks fall hardest on those with the least power to contest them. Algorithms trained on skewed datasets reproduce and often amplify existing social hierarchies. An AI credit-scoring model that penalises women for interrupted employment histories, or a resource allocation tool that under-counts marginalised populations absent from formal records, does not represent neutral technology; it encodes discrimination at scale.There are encouraging signs that lawmakers are responding. India’s AI Governance Guidelines, released in November 2025, anchor fairness and equity as core principles, with special emphasis on protecting marginalised communities. The Digital Personal Data Protection Rules, notified around the same time, bring 800 million internet users under a formal privacy framework for the first time. A Private Member’s Bill introduced in the Lok Sabha in December 2025 goes further, proposing mandatory bias audits, restrictions on AI in surveillance and employment screening, and statutory penalties. Not yet a law, but it considers accountability in the legislature.The labour question cannot be deferred, particularly because AI is expected to displace routine tasks across the economy, and the effects will be uneven. Workers in low-wage, repetitive roles, overwhelmingly from marginalised backgrounds, are most exposed. A framework that invests in AI infrastructure without investing equally in reskilling, social protection, and equitable access to productivity gains would trade short-term efficiency for long-term fracture.India’s AI Governance Guidelines need to consider a principle-based, techno-legal approach over a top-down regulatory regime, i.e., embedding accountability into system design rather than policing it after deployment. Proposed institutions like an AI Governance Group, a Technology and Policy Expert Committee, and an AI Safety Institute, are a positive way forward to the architecture. Algorithmic audits, impact assessments, and mandatory incident reporting need to be operationalised. Civil society and independent researchers need access to audit trails, not just regulators.Crucially, none of this replaces human judgment. The most thoughtful AI deployments in governance position the technology as a support layer for officials, not a substitute for democratic deliberation. When a model flags a district as high-risk for malnutrition, a person still decides the response. The technology expands what is knowable; it cannot, and should not, determine what is done.India’s choice is not between AI and no AI. It is between AI adopted thoughtfully, with benefits channelled through the same public infrastructure logic that animated Aadhaar and UPI, and AI that drifts toward serving those who already hold data, capital, and institutional voice. The Madhya Pradesh experiment, the November guidelines, the parliamentary debate on accountability: These are early chapters of a story still being written. In a fast-changing world, this transition is a necessity for optimising resource efficiency, improving public services, and staying ahead of mounting social and economic challenges. But it must be done with caution: With robust safeguards, inclusive design, and the citizen, not the algorithm, at the centre.(The views expressed are personal)This article is authored by Chetana Chaudhuri, fellow, National Council of Applied Economic Research (NCAER), and Raj Kumar Banerjee, founder and managing director, Isourse Technologies.