🗞️ Why in News 88 nations endorsed the New Delhi Declaration on AI Impact at the India AI Impact Summit 2026 in New Delhi on February 22, 2026. Built on seven “Chakra” pillars including democratising AI resources and ensuring trusted AI, the declaration positions India as a key architect of inclusive global AI governance — distinct from the Western-led safety-first approach of the Bletchley Park and Seoul summits.
Context: The Global AI Governance Race
Since ChatGPT’s launch in November 2022, artificial intelligence has dominated global policy discourse. Advanced AI systems have moved from research labs to commercial deployment at scale, raising pressing questions about safety, misuse, economic disruption, and geopolitical competition. The international community has struggled to agree on a governance framework.
Three major diplomatic initiatives preceded New Delhi:
- Bletchley Park AI Safety Summit (UK, November 2023): Focused on frontier AI safety risks; produced the Bletchley Declaration signed by 28 countries; launched AI Safety Institutes in the UK and USA
- Seoul AI Summit (South Korea, May 2024): Extended safety commitments; included 16 AI companies in frontier AI pledges; focused on red-lines for “highly capable AI”
- Paris AI Action Summit (France, February 2025): Broader participation; focused on AI for public interest; contested US position on binding regulation
India’s approach is deliberately different: less focused on existential AI risk (the concern of wealthy countries with frontier AI labs) and more on equitable access, economic inclusion, and democratic diffusion of AI capabilities.
What the New Delhi Declaration Says
The Declaration’s seven Chakra pillars are:
| Pillar | Core Principle |
|---|---|
| 1. Democratising AI Resources | Foundational AI tools, compute, and data should be accessible to all nations |
| 2. Economic Growth and Social Good | AI development must create broad-based benefit, not concentrate wealth |
| 3. Secure and Trusted AI | Safety standards, transparency, accountability — but through capacity building, not only regulation |
| 4. AI for Science | Accelerate scientific research using AI, especially for global challenges (climate, health, food) |
| 5. Social Empowerment | Bridge the AI divide; ensure marginalised communities benefit |
| 6. Human Capital Development | AI workforce training and education, especially for developing nations |
| 7. Resilient and Innovative AI Systems | Infrastructure, standards, and innovation ecosystems |
The Charter for the Democratic Diffusion of AI operationalises Pillar 1: it proposes shared access to AI compute infrastructure, open foundational models, and multilingual AI development. This addresses a structural asymmetry: the USA and China have dominant AI compute clusters and proprietary foundation models, while the rest of the world depends on their APIs.
India’s Strategic AI Positioning
India’s approach to global AI governance reflects its broader diplomatic identity as a Global South leader and bridge between blocs. The New Delhi Declaration’s voluntary, non-binding character is consistent with India’s preference for sovereignty-preserving frameworks (as seen in its positions on data governance, digital trade, and cyber norms).
Key elements of India’s AI strategic calculation:
1. India AI Mission: India approved the India AI Mission in 2024 with a budget of Rs 10,372 crore over 5 years, focusing on:
- Public AI compute infrastructure (shared GPU clusters for research and startups)
- AI in agriculture, health, and public service delivery
- Indian language models (India has 22 scheduled languages and over 100 languages with significant speaker populations)
- AI safety and trustworthiness standards aligned with India’s values
2. IndiaAI initiative: The government has been building an IndiaAI platform under MeitY (Ministry of Electronics and Information Technology) to coordinate AI policy, facilitate public-private partnerships, and publish AI readiness assessments.
3. Compute sovereignty: India’s biggest strategic AI vulnerability is compute — the specialised GPUs and TPUs used for training and running large AI models are manufactured by NVIDIA, AMD, Google, and a handful of others, overwhelmingly in the USA and Taiwan. The New Delhi Declaration’s focus on democratising compute access directly reflects India’s interest in reducing this dependence.
India vs. EU vs. USA Approaches to AI Governance
| Dimension | EU AI Act (2024) | USA Executive Orders | India New Delhi Declaration |
|---|---|---|---|
| Binding? | Yes — legally binding regulation | Partially (executive orders) | No — voluntary |
| Focus | Risk categorisation; prohibitions; compliance | Safety, security, innovation | Access, inclusion, economic growth |
| Who it protects | EU citizens and markets | US interests and innovation | Global South, developing nations |
| Enforcement | National AI authorities; fines up to €35M | Federal agency guidance | Multistakeholder dialogue |
| AI liability | High-risk AI must be audited | Emerging framework | Not addressed |
India’s approach is closest to the Global Partnership on AI (GPAI) model — voluntary, multistakeholder, and consensus-based. It avoids the regulatory rigidity of the EU (which many developing countries cannot afford to comply with) and the US focus on frontier AI company self-regulation.
Implications for UPSC
The New Delhi AI Declaration is significant for multiple GS papers:
GS-3 (S&T): India AI Mission, compute infrastructure, foundational AI models, AI safety concepts, CERT-In and AI security
GS-2 (IR + Governance): India’s AI diplomacy, multilateral governance frameworks, India’s Global South leadership, sovereignty vs. rules-based global governance tension, digital colonialism concerns
GS-2 (Polity): MeitY’s role, data governance (DPDP Act 2023 intersection), regulatory capacity building
UPSC Relevance
Prelims: New Delhi Declaration on AI Impact (Feb 22, 2026), India AI Impact Summit 2026, 7 Chakra pillars, Charter for Democratic Diffusion of AI, India AI Mission (Rs 10,372 crore, 2024), MeitY, GPAI (Global Partnership on AI), Bletchley Declaration (2023), EU AI Act (2024), IndiaAI platform, Vasudhaiva Kutumbakam. Mains GS-2: AI governance; India’s multilateral diplomacy; digital sovereignty vs. global norms. GS-3: Artificial intelligence; national AI missions; India’s technology strategy; compute infrastructure; AI safety frameworks. Interview: “India promotes a voluntary, sovereignty-preserving approach to AI governance while the EU has adopted binding regulation. Which approach better serves the global public interest — and what does India’s choice reveal about its foreign policy priorities?”
📌 Facts Corner — Knowledgepedia
New Delhi AI Declaration 2026:
- Summit: India AI Impact Summit 2026 | Date: February 22, 2026 | Location: New Delhi
- Signatories: 88 countries (USA, UK, China, Brazil, Australia, Canada, Austria, Switzerland + 80 more)
- Structure: 7 pillars (“Chakras”) + Charter for Democratic Diffusion of AI
- Nature: Voluntary, non-binding | Philosophical basis: Vasudhaiva Kutumbakam
India AI Mission:
- Approved: 2024 (Union Cabinet) | Budget: Rs 10,372 crore | Duration: 5 years
- Nodal ministry: MeitY (Ministry of Electronics and Information Technology)
- Components: Public compute infrastructure; Indian language models; AI startups; AI safety
- Platform: IndiaAI (centralized portal for AI policy + resources)
Global AI Governance Milestones:
- Nov 2023: Bletchley Park AI Safety Summit (UK) | 28 countries | Bletchley Declaration
- May 2024: Seoul AI Summit (South Korea) | Frontier AI safety pledges (16 companies)
- Feb 2025: Paris AI Action Summit (France) | AI for public interest
- Feb 2026: New Delhi AI Impact Summit (India) | 88 nations | New Delhi Declaration
EU AI Act (2024):
- World’s first legally binding AI regulation
- Risk-based framework: Unacceptable risk (banned) | High risk (audited) | Limited risk (transparency) | Minimal risk (free)
- Unacceptable risk examples: Social scoring by govts, real-time biometric surveillance, subliminal manipulation
- Fines: Up to €35 million or 7% of global turnover
- Enforced by: National AI authorities in each EU member state
GPAI (Global Partnership on AI):
- Founded: 2020 by G7 countries + EU | India joined: 2020
- Secretariat: OECD, Paris
- Focus: Responsible AI development through multistakeholder research
- India assumed GPAI Presidency: 2023
Compute and AI:
- GPU (Graphics Processing Unit): Key hardware for AI training (NVIDIA H100 most used)
- LLM (Large Language Model): ChatGPT, Gemini, Claude, Llama — core AI products
- Foundation model: Pre-trained model used as base for many AI applications
- Compute access gap: USA/China have dominant GPU clusters; most countries rent access via API