India - National AI Strategy

Overview

The National Strategy for Artificial Intelligence (#AIforAll), published by NITI Aayog in June 2018, is a strategic discussion paper that proposes a roadmap for leveraging AI to advance inclusive economic growth and social development in India. The document argues that India can develop a unique brand of AI—"#AIforAll"—by focusing AI efforts on public good use cases across healthcare, agriculture, education, smart cities and mobility. It explains why these sectors offer high social return on investment and proposes institutional, technical and programmatic measures to scale AI solutions nationally and replicate them in similar emerging economies. The Strategy is not a binding law but a policy blueprint designed to coordinate government, industry and academic actors to catalyse AI research, build public data ecosystems, and promote responsible deployment. The full discussion paper and subsequent consolidated versions are available from the NITI Aayog repository (see Sources).

Definitions

For clarity and operational use, the Strategy defines key terms such as "Artificial Intelligence" (a set of techniques enabling machines to mimic cognitive functions including learning, reasoning and perception), "Machine Learning" (a subset of AI that learns patterns from data), "Centres of Research Excellence (CORE)" (institutions focused on fundamental research), and "International Centres for Transformational AI (ICTAI)" (applied, domain-specific labs partnering with industry). The paper also distinguishes between "public-good AI"—applications that address societal needs—and "commercial AI"—market-driven products—while stressing overlaps and mutual dependence. Data ecosystem terminology is clarified: "curated public datasets," "data registries," and "data custodians" are introduced to describe roles for data access, anonymization and governance.

Governance and Institutional Framework

The Strategy recommends a multi-layered governance approach. At the centre is an AI oversight and coordination body—proposed to be hosted or convened by NITI Aayog—to set national standards, benchmarks and guidelines and to coordinate cross-sectoral pilots with sectoral regulators. The paper proposes establishing COREs and ICTAIs as institutional anchors: COREs to advance foundational research and capacity-building, and ICTAIs to develop applied solutions with industry and public sector partners. It recommends that the proposed oversight body work with ministries such as the Ministry of Electronics and Information Technology (MeitY), health, agriculture and education departments, state governments and sectoral regulators to harmonize data access, procurement standards and certification practices. The Strategy also calls for public funding mechanisms to support demonstrator projects, seed grants for startups, and skilling initiatives to expand India’s AI workforce.

Key Focus Areas

The paper identifies five initial priority sectors where AI can accelerate inclusive outcomes: (1) Healthcare: scaling diagnostics (e.g., radiology/ pathology imaging), population health analytics and early-detection systems; (2) Agriculture: yield prediction, crop disease detection, and supply-chain optimization to raise farmer incomes and reduce losses; (3) Education: personalized learning platforms, adaptive assessments and teacher-support tools to improve learning outcomes at scale; (4) Smart Cities & Infrastructure: predictive maintenance, energy optimisation, waste management and citizen services to manage rapid urbanisation; and (5) Smart Mobility & Transportation: traffic prediction, route optimization and safety systems to reduce congestion and accidents. For each sector the Strategy lists concrete use cases, potential data sources, stakeholder roles, and implementation challenges including data quality, interoperability and model robustness. Cross-sector themes include developing labeled datasets, establishing ethics-by-design, and ensuring human-in-the-loop oversight where decisions affect rights or safety.

Implementation Framework

Implementation is structured around three execution strategies: (A) rapid experimental pilots and Proof-of-Concepts (PoCs) across the priority sectors to demonstrate value and create adoption templates; (B) building the AI ecosystem through investments in CORE and ICTAI, public datasets, cloud and compute infrastructure, and national skilling initiatives; and (C) partnerships and international collaboration to bring best practices and standards to India. The document proposes governance measures for procurement—encouraging public agencies to require explainability, audit trails and conformance with standards in AI procurements—and suggests funding instruments (grants, challenge prizes and co-funding with industry). It highlights the need for evaluation sandboxes and testbeds to safely assess AI systems before scale-up.

Monitoring and Evaluation

The Strategy prescribes a monitoring and evaluation (M&E) approach that combines technical metrics (accuracy, robustness, fairness tests), deployment KPIs (uptake rates, service coverage), and social impact measures (access, affordability and equity outcomes). It advocates use of standardized benchmarking datasets and sectoral performance indicators, and recommends periodic public reporting on pilot outcomes. The oversight body would collate results, identify high-impact projects for scale-up, and maintain a public registry of validated solutions. The Strategy also encourages independent third-party evaluation for higher-risk use cases to improve trust and transparency.

Penalties, Liability, and Appeals

As a strategy document, #AIforAll does not itself prescribe statutory penalties or enforcement regimes; instead it recommends that sectoral regulators integrate AI-specific safety, liability and accountability considerations into existing regulatory frameworks. The paper calls for clear allocation of responsibilities among system developers, deployers and procurers, and for grievance redress mechanisms where automated decisions affect rights. It proposes that procurement contracts include audit and liability clauses and that oversight mechanisms require incident reporting and corrective action plans. Where necessary, sectoral laws (for example in healthcare or transportation) would define civil or administrative liabilities for harms resulting from AI-enabled systems.

Relationship to Other Instruments

The Strategy is positioned as a coordinating policy that complements sectoral laws, data protection initiatives and national digital programmes. It links to ongoing work on national data governance, digital health initiatives and skill-development schemes. The document anticipates close alignment with a national data protection framework (data anonymization, consent standards) and advises regulators to update sector-specific rules—such as clinical trials, agricultural extension norms and transport safety standards—to account for AI. It also explicitly encourages harmonization with procurement rules, open-data policies and research funding mechanisms to ensure joined-up implementation.

International Alignment

The Strategy recommends active international engagement to adopt global best practices and standards, participate in multilateral AI governance fora, and promote India’s approach in other emerging economies. It suggests collaboration on research, dataset sharing (subject to privacy safeguards), and standards development. The paper emphasizes that India’s "AI for All" brand should reflect priorities of inclusion, affordability and replicability for other developing nations, balancing innovation with safeguards around privacy, safety and human rights. NITI Aayog recommends representing India in international dialogues and leveraging bilateral and multilateral partnerships for capacity building.

Implementation Timeline

Phase	Action	Indicative Timeline
Phase 1 (Pilot)	Identify and run PoCs in priority sectors; establish data registries and testbeds	0–12 months
Phase 2 (Scale)	Set up CORE and ICTAI centres; expand successful pilots; develop procurement guidelines	12–36 months
Phase 3 (Sustain)	National roll-out of validated solutions, long-term funding and international partnerships	36+ months

Compliance Checklist

Requirement	Checklist
Data readiness	Has required curated datasets been identified and anonymized? Are data custodians assigned?
Governance	Is there a designated oversight/coordination lead and sectoral regulator engagement?
Safety & Evaluation	Are testbeds/benchmarks and third-party evaluation plans in place?
Procurement	Do procurement contracts include explainability, audit and liability clauses?
Public interest	Have equity and access impacts been assessed and mitigation measures defined?

Sources and References

Source	Type
National Strategy for Artificial Intelligence (#AIforAll) - NITI Aayog (discussion paper / strategy, June 2018)	Primary Source
Responsible AI: NITI Aayog (Responsible AI principles, 2021)	Primary Source