India’s AI Policy Needs to Get Global-Scale Ready

UPSC Syllabus Topic: GS Paper 3 –Awareness in the fields of IT, Space, Computers, robotics, nano-technology, bio-technology.

Introduction

India’s AI governance framework is gaining attention for its responsible and inclusive approach, built on digital public infrastructure (DPI) and voluntary guardrails, not heavy regulation. This has helped scale AI use across services. However, the core concern remains unresolved. India’s policy support is still stronger for AI deployment than for foundational AI creation, which raises long-term issues of sovereignty, resilience, and global competitiveness.

Current status of India’s AI ecosystem

1. Strong digital public infrastructure base: India’s digital public infrastructure gives AI developers a rare structural advantage. Platforms such as Aadhaar, UPI, DigiLocker, Bhashini, and DEPA allow fast scaling of identity, payments, data sharing, and service delivery. This integrated national architecture supports real-world AI use at population scale, which few countries can replicate.
2. Rapid growth in AI adoption and industry size: India’s technology sector is projected to cross USD 280 billion in annual revenue. Over 6 million people are employed in the tech and AI ecosystem. Around 87% of enterprises actively use AI solutions, with strong adoption in BFSI, healthcare, manufacturing, retail, and automotive sectors.
3. Expanding startup and developer ecosystem: India hosts around 1.8 lakh startups, and nearly 89% of new startups launched last year used AI in their products or services. The country has 1,800+ Global Capability Centres, including 500+ AI-focused centres. India is also the second-largest contributor to AI projects on GitHub, showing strong developer participation.

4. Global recognition and competitiveness: India ranks among the top four countries in AI skills and policy readiness and stands third globally in AI competitiveness. This reflects talent depth, research output, startup activity, and digital infrastructure, but the strength remains skewed towards deployment rather than core model building.

Challenges and Concerns Related to Pillars of India’s AI Strategy

1. Foundational Gap: While AI applications are growing, support for building core AI models remains limited. Most developers rely on fine-tuning existing open models instead of training original ones. This creates a gap between application strength and core technological ownership.
2. Data Uncertainty: There is no legal clarity on whether publicly available data can be used for AI training. The Copyright Act is not updated and there is no text-and-data mining exemption, so developers cannot be sure they are within legal limits.
3. Liability Doubts: The policy does not clearly define responsibility when AI systems cause harm. It is unclear whether liability lies with the developer, deployer, or platform. This uncertainty increases risk, especially for small firms and those working in finance or healthcare.
4. Compute Bottleneck: India has strong academic talent in AI, but researchers lack easy access to compute resources. The AIRAWAT initiative is promising but remains opaque. Access rules are unclear and approvals are slow, which limits experimentation and delays research progress.

For detailed information on AI Supercomputer ‘AIRAWAT read this article here

5. Foreign Dependence: Low domestic model building increases reliance on foreign models. This ties India to external licensing terms, design choices, and future support decisions, creating long-term dependency.
6. Sovereignty Risk: When public services and user experiences depend on systems built elsewhere, control reduces. This can weaken resilience and competitiveness over time.

Initiatives taken to strengthen India’s AI strategy

1. IndiaAI Mission: The government launched the IndiaAI Mission in March 2024 with over ₹10,300 crore to build a strong indigenous AI ecosystem. It aims to democratise AI technology, improve data quality, and boost competitiveness.
2. INDIAai National Portal: INDIAai is India’s national AI portal supporting knowledge sharing, research insights, industry news and resources to connect stakeholders across the ecosystem.
3. AI Governance Guidelines: Government has rolled out governance guidelines under IndiaAI that aim to make AI safe, inclusive and pro-innovation without heavy regulation that stifles growth.
4. Centres of Excellence (CoEs):The government set up three CoEs in Healthcare, Agriculture, and Sustainable Cities, and announced a fourth CoE for Education in Budget 2025, to support collaborative and scalable AI innovation.
5. Sarvam AI and BharatGen Models: Initiatives like Sarvam AI (for smarter public services) and BharatGen AI (multilingual, multimodal model) focus on homegrown capabilities that reflect India’s linguistic and cultural diversity.
6. AI Impact Summit 2026: India will host the AI Impact Summit to showcase its AI capabilities, encourage innovation and build international collaborations.

What should be done?

1. Legal clarity: A targeted step is to confirm that training on publicly available data for AI research is legal in India. This would lower legal risk and unlock broader experimentation in academia and startups.
2. Create predictable safe harbours: Safe harbours, like those used for internet intermediaries, can protect developers from automatic liability for third-party misuse. Liability should be proportional, predictable, and linked to real control.
3. Make AIRAWAT truly usable: Publish clear access norms, create simple onboarding, and offer shared compute clusters with minimal approvals. Researchers and small firms should be able to start training jobs without long delays.
4. Run regulator-backed sandboxes: Structured sandboxes in sectors like finance and healthcare should be run with regulators and backed by legal guidance. This supports safe testing without informal or ad-hoc experimentation.
5. Light certification: A simple certification check for fairness, transparency, and robustness can create a clear incentive to build responsibly. If a model meets baseline tests, it should plug into DPI use cases by default.
6. Learn from enabling states:

• The UAE launched Falcon, an open-source language model, with clear government backing and high visibility.
• Singapore is building rules that push explainability and give users a path for redress.
• The EU, even with stricter regulation, offers clearer certainty and research carve-outs.
• The US still gives developers room to build under a broad fair use approach.
• The shared lesson is simple: policy clarity and trust rules can enable builders, not just control them.

Conclusion

India has a strong edge in AI deployment due to DPI, talent, and market scale. To become global-scale ready, it must close gaps in core model building, clarify data legality, ensure predictable liability, and make AIRAWAT practically accessible. These targeted steps can reduce dependency and position India as a sovereign and trusted AI builder.

Question for Practice

Discuss the key strengths and main gaps in India’s AI strategy that must be fixed to achieve global-scale AI capability.

Source: Businessline