85 countries, 3 bodies sign New Delhi Declaration for equitable AI at Summit

New Delhi Declaration for Equitable AI — UPSC Study Note

1. At a Glance

• 85–92 countries and 3 international organisations signed the New Delhi Declaration on AI Impact at the AI Impact Summit 2026, held in New Delhi on 18–19 February 2026. [S1][S2]
• India positioned itself as a global convener on AI governance, bridging the Global North–South divide on AI access, safety, and equity. [S1]
• Guiding principle: "Sarvajan Hitaya, Sarvajan Sukhaya" (Welfare for all, Happiness for all) — India's framing of AI as a developmental, not merely technological, issue. [S1][S2]
• UPSC relevance: GS-II (International relations, multilateral diplomacy), GS-III (Technology, AI governance, economic development). Active, recurring exam theme since AI governance featured in G20 2023 New Delhi Summit.

2. Why in the News

• 18–19 February 2026: India hosted the AI Impact Summit 2026 in New Delhi; Declaration adopted with 85 countries at conclusion (later expanded to 92 as additional countries joined in subsequent days). [S1][S3]
• The Summit followed the Paris AI Action Summit (February 2025), where the U.S. notably refused to endorse the Paris communiqué on AI safety — the New Delhi Declaration deliberately incorporated softer, non-binding language on AI safety and trust to achieve U.S. and China endorsement. [S2][S4]
• Union Minister Piyush Goyal presided; notable attendees included U.S. Ambassador Sergio Gor, Bhutan PM Tshering Tobgay, and OpenAI CEO Sam Altman. [S4]

3. Background & Evolution

• 2023 G20 New Delhi Summit: India first embedded AI governance principles in a G20 Leaders' Declaration; precedent of consensus through non-binding language established. [S2]
• November 2023: UK-hosted Bletchley Park AI Safety Summit — first major multilateral AI summit; focused heavily on frontier AI risks. [Background knowledge, contextualised by S2]
• May 2024: Seoul AI Summit — second in the safety-focused series. [Background, contextualised by S2]
• February 2025: Paris AI Action Summit — France hosted; U.S. and U.K. declined to sign final statement, fracturing Western consensus on AI safety framing. [S2]
• February 2026: India's AI Impact Summit 2026 — deliberate pivot from safety-first framing toward equitable access and development-led AI governance, positioning India as champion of the Global South. [S1][S2]

4. Core Static Facts

5. Multi-Dimensional Analysis

Economic

• USD 200 billion+ in AI investments expected across infrastructure, foundation models, hardware, and applications as a direct Summit outcome. [S1]
• ILO–India Equitable AI Transition Playbook addresses worker reskilling — positions AI as an employment opportunity, not only a displacement risk. [S1]
• Global AI Impact Commons features 80+ impact stories across 30+ countries, enabling replication of AI use-cases for economic growth in developing nations. [S2]

Geopolitical / Strategic

• USA and China both signing is diplomatically significant given bilateral tensions; India's non-aligned convener role echoes its G20 "Vishwabandhu" branding. [S1][S4]
• Deliberate contrast with Paris Summit (2025) where U.S.–UK refused to sign: New Delhi Declaration adopted softer AI safety language to achieve broader consensus. [S2][S4]
• Declaration cements India's claim to AI governance leadership for the Global South, complementing its Digital Public Infrastructure (DPI) diplomacy.
• The three international organisations signing extend the Declaration's legitimacy beyond state actors into multilateral institutions. [S1]

Social / Equity

• Core framing: AI benefits must be "equitably shared across humanity" — explicit Global South orientation. [S1]
• Charter for Democratic Diffusion targets "affordable access to foundational AI resources" and "locally relevant innovation" — addresses compute and data inequality. [S2]
• Trusted AI Commons provides tools, benchmarks, and best practices, lowering barriers for capacity-constrained nations. [S2]

Scientific / Technological

• Charter for the Democratic Diffusion of AI: voluntary framework promoting access to foundational AI resources, local innovation ecosystems, and resilient AI infrastructure. [S2][S4]
• Global AI Impact Commons: open repository of scalable AI use-cases across sectors and countries. [S2]
• Trusted AI Commons: shared pool of AI safety tools, testing benchmarks, and secure development practices. [S2]

Ethical / Governance

• All commitments are voluntary and non-binding — mirrors G20 2023 approach; critics note this limits enforceability. [S1][S4]
• Declaration "does mention AI safety and trust" but commitments "lean heavily on knowledge sharing" — a deliberate geopolitical balance between safety-first (U.S./EU) and growth-first (Global South) camps. [S4]
• Absence of a binding enforcement mechanism raises questions about accountability (a likely Mains discussion point).

Administrative

• Post-Summit, additional countries continued to sign (signatories expanded from 85 → 92+), suggesting ongoing diplomatic effort beyond the Summit dates. [S3]
• India's MeitY led domestic coordination; MEA managed multilateral diplomatic outreach. [S1]

6. Recent Developments (Last 12–18 Months)

• February 2025: Paris AI Action Summit — U.S. and U.K. declined to sign final communiqué; fractured Western bloc on AI safety. [S2]
• 18–19 February 2026: AI Impact Summit 2026 held in New Delhi. [S1]
• 21–22 February 2026: Declaration signing extended by one day to expand signatory list from initial 85; final figure at announcement: 85 countries + 3 international bodies. [S1][S4]
• Post-22 February 2026: Signatories rose to 89, then 91, then 92 as Bangladesh, Costa Rica, Guatemala and others joined. [S3]
• ILO co-release: Equitable AI Transition Playbook published alongside the Declaration. [S1]
• USD 200 billion+ investment pipeline announced as Summit outcome. [S1]

7. Prelims Hooks

1. The AI Impact Summit 2026 was held in New Delhi on 18–19 February 2026. [S1]
2. The New Delhi Declaration was initially signed by 85 countries and 3 international organisations. [S1]
3. The guiding principle of the Declaration: "Sarvajan Hitaya, Sarvajan Sukhaya" (Welfare for all, Happiness for all). [S1]
4. All commitments in the New Delhi Declaration are voluntary and non-binding. [S1][S4]
5. The Charter for the Democratic Diffusion of AI was supported by 22 countries and international institutions. [S2]
6. Global AI Impact Commons features 80+ impact stories across 30+ countries. [S2]
7. The Declaration is approximately 900 words in length. [S4]
8. Both the United States and China endorsed the New Delhi Declaration. [S4]
9. India co-released the Equitable AI Transition Playbook with the ILO at the Summit. [S1]
10. The Summit was presided over by Union Minister Piyush Goyal. [S4]
11. Signatories later rose to 91–92 with Bangladesh, Costa Rica, and Guatemala among late joiners. [S3]
12. The Trusted AI Commons aims to share tools, benchmarks, and best practices for secure AI development. [S2]
13. 100+ countries participated in proceedings; 20+ Heads of State and 60+ Ministers attended. [S1]
14. The AI Impact Summit 2026 followed the Paris AI Action Summit (February 2025) in the global AI governance calendar. [S2]
15. AI-related investment commitments announced at Summit: USD 200 billion+. [S1]

8. Mains Relevance

GS Paper Mapping: - GS-II: International relations — India's multilateral diplomacy; global governance of emerging technologies; India and the Global South. - GS-III: Science & Technology — Artificial Intelligence governance, policy frameworks; Digital economy; Effects of technology on employment.

Specific Syllabus Headings: - GS-II: Bilateral, regional, and global groupings and agreements involving India; Important international institutions. - GS-III: Awareness in the field of IT, Space, Computers, Robotics, Nano-technology, Bio-technology; Effects of liberalization on the economy.

Plausible Mains Question Stems: 1. "The New Delhi Declaration on AI Impact represents a paradigm shift from a safety-first to an equity-first approach to AI governance. Critically examine India's role in reshaping the global AI governance architecture." (GS-II) 2. "Voluntary and non-binding frameworks have been the preferred mode for international technology governance. Analyse the limitations of this approach with reference to the New Delhi Declaration on AI Impact, 2026." (GS-II/GS-III) 3. "How does the concept of 'Democratic Diffusion of AI' address the concerns of developing nations in the global AI order? What are the structural barriers to its realisation?" (GS-III)

9. Related Topics to Study Next

10. Common Errors / Trap Areas

1. Confusing signatory count: Initial signing = 85 countries + 3 bodies; rose to 91–92 later. Prelims may test either figure — note the date context.
2. Host Ministry: Summit hosted by MeitY (not MEA, not NITI Aayog). MEA managed diplomatic outreach but MeitY was the nodal ministry.
3. Non-binding nature: Aspirants often treat "signed" as equivalent to "legally binding." The New Delhi Declaration, like the G20 AI principles, is entirely voluntary and non-binding.
4. Confusing the three frameworks: (a) Charter for Democratic Diffusion = policy/access framework; (b) Global AI Impact Commons = repository of AI use-cases; (c) Trusted AI Commons = safety tools/benchmarks. These are three distinct instruments, not synonyms.
5. Paris Summit confusion: The Paris AI Action Summit (Feb 2025) is a different event from the New Delhi Summit (Feb 2026); U.S./U.K. signed New Delhi but not Paris — aspirants often invert this.

11. Sources

• [S1] AI Impact Summit 2026 Concludes with Adoption of New Delhi Declaration — https://www.pib.gov.in/PressReleasePage.aspx?PRID=2231208 — (Tier 1)
• [S2] India AI Impact Summit 2026: Landmark Global Declaration and Major AI Investment Commitments — https://www.pib.gov.in/PressReleasePage.aspx?PRID=2234343 — (Tier 1)
• [S3] More countries join the New Delhi Declaration on AI Impact — https://www.pib.gov.in/PressReleasePage.aspx?PRID=2232005 — (Tier 1)
• [S4] 85 countries, 3 bodies sign New Delhi Declaration for equitable AI at Summit (Aroon Deep) — The Hindu, 22 February 2026 — https://www.thehindu.com/todays-paper/2026-02-22/th_international/articleGSBFKDQ5S-13608250.ece — (Tier 4)