Democratising AI in India

The ₹6 Million Opportunity: How India is Democratising AI

Over 6 million Indians are currently employed in artificial intelligence (AI)-related sectors, a staggering figure that underscores the transformative economic potential of AI in India. Yet, this number is dwarfed by the untapped possibilities in healthcare, agriculture, education, and governance, where true democratisation of AI—ensuring fair and wide access to AI tools, data, and infrastructure—could fundamentally alter development trajectories. The India–AI Impact Summit 2026, convened this week, aims to formalise frameworks for inclusive AI adoption both domestically and across the Global South. But beneath the optimism lie hard questions about equity, infrastructure, and regulation.

India’s Model: Breaking from the Normal Tech Paradigm

What India attempts with AI is distinctive. Unlike in advanced economies, where AI ecosystems often operate within elite silos of global tech corporations, India is championing what may be called "baseline democratisation." This is evident in initiatives like SOAR (Skilling for AI Readiness), which introduces AI fundamentals and ethics to students as young as 12, and the IndiaAI Mission, which is funding 500 PhDs and setting up AI labs even in Tier-2 and Tier-3 cities. The Digital Personal Data Protection Act (2023) is a further cornerstone, aiming to build trust in how personal data fuels AI systems. More revolutions than reforms, these shifts place accessibility and inclusivity—the Global South’s defining needs—at the centre.

Contrast this with the US approach, where corporate-led breakthroughs in frontier AI models like GPT-4 remain largely confined to proprietary platforms, or China’s opaque deployment of AI within state-owned domains. India’s thrust on openness—exemplified by AIKosh, a repository of Indian datasets and reusable models—offers a blueprint for others seeking to democratise data while maintaining safeguards.

The Infrastructure Deficit: Who Pays for Democratisation?

The real hurdle lies not in intent but in capacity. Access to computing power, high-quality datasets, and skilled manpower—vital ingredients of AI ecosystems—remains uneven within India. The GI Cloud (MeghRaj), India's primary digital infrastructure backbone, aims to reduce cost and technical barriers through on-demand, pay-as-you-use cloud access for government projects. Yet rural and underserved regions still find themselves with inadequate infra to leverage these tools. Despite MeghRaj’s theoretical scalability, challenges like patchy broadband, limited last-mile connectivity, and urban-rural digital divides dilute its potential reach.

Another glaring gap is funding. The promised benefits of democratised AI hinge on equitable distribution of resources, which MeitY supports through initiatives like Centres of Excellence, but the actual outlay has stagnated. The publicly available budget indicates MeitY’s allocation to AI-driven projects was ₹4,000 crore in 2025, a figure that pales in comparison to the US’s $1.2 billion annual federal AI research budget. The irony here is that democratisation is no less expensive than centralised AI deployment—if anything, it requires more distributed and resilient infrastructure investment.

AI Governance: Widening Definitions, Narrowing Accountability

Data privacy remains both the bedrock and the Achilles’ heel of India’s AI ambitions. The Digital Personal Data Protection Act (2023) is framed as an enabler of trust, but gaps in oversight mechanisms persist. For instance, while the legislation mandates purpose-specific data collection and imposes steep penalties for breaches, India’s track record in regulatory enforcement—particularly in data-heavy sectors like telecom—offers limited reassurance. Reports of systemic bias in AI implementations also raise concerns. Earlier this year, flaw-ridden AI credit-scoring models used by banking institutions disproportionately penalised applicants from low-income castes. Without rigorous pre-deployment ethical audits and continuous oversight, there’s a risk that AI democratisation amplifies, rather than corrects, existing socio-economic disparities.

Another blind spot is the blurred line between open access and privatisation. Global tech companies, riding on government subsidies for AI research, increasingly dictate how AI systems evolve. While the India–AI Impact Summit touted shared resource pools with countries like Kenya and Egypt, little was said about how commercial confidentiality agreements would coexist with open innovation goals.

Global Comparisons: India and South Korea Diverge

Consider South Korea's approach, where the National AI Strategy since 2019 has channelled $2 billion annually to democratise core resources—focusing especially on high-speed 5G infrastructure and expanding nationwide AI hubs. By 2023, 44% of South Korean small- and medium-scale enterprises were actively deploying AI solutions under co-financed public-private ventures. Compare this to India's fragmented funding and implementation, where broadband penetration in rural India is still below 40% and AI hubs are concentrated in urban IT clusters. While South Korea leveraged its compact urban geography for targeted resource amplification, India’s scale—both an advantage and a constraint—demands a decentralised but standardised delivery model. India’s efforts to counteract these limitations, such as by training educators in AI fundamentals under YUVAi, remain nascent in reach.

Uncomfortable Questions That Remain

If inclusivity is at the heart of India’s AI vision, one uncomfortable question persists: who benefits first, and who benefits the most? With nearly 90% of India’s startups leveraging AI for applications like med-tech and agritech, how does resource allocation balance between the startup ecosystem and public services? Much of the visible AI progress remains biased toward revenue-generating sectors rather than public welfare initiatives like smart rural governance. Additionally, while AIKosh promotes indigenous datasets, the dominance of English-language data risks compromising inclusivity in multilingual applications critical for rural service delivery.

Another unresolved tension is India’s stance on global AI ethics frameworks. As co-chair of the Democratising AI Resources Working Group, India champions accessibility for Global South nations. But criticism—largely warranted—points to India’s limited domestic success in operationalising its own ethical AI goals, such as eliminating algorithmic transparency deficits. Without setting its house in order, India’s credibility in setting global standards may be at risk.

Conclusion: Walking the Talk on Inclusive AI

The underlying premise of democratising AI is powerful: decentralising tools that historically remain monopolised. India’s early steps, from SOAR in schools to the IndiaAI Mission in higher education, hold promise for systemic inclusion. Yet the structural weaknesses—gaps between legislative framework and enforcement, underfunded rural infra, and a limited multilingual AI ecosystem—must be addressed before AI’s full democratising potential can be unleashed. Detached from India’s development narrative, AI would risk being another technology elite capture story. Embedded within it, however, AI could redefine how innovation serves public purpose.

Practice Questions for UPSC

Prelims Practice Questions

Frequently Asked Questions