AI for India’s Next Billion: Intergenerational Insights for Inclusive and Future-Ready Growth

Detailed Summary

The session opens with Amandeep Singh Gill (UN) thanking the Government of India and the United Nations‑University (UW) for co‑hosting the summit. He introduces accountability as the “easy question” that underpins trustworthy AI governance.

Key points presented by Gill

• Diffuse accountability can either enable a flexible, adaptive AI ecosystem or become a “recipe for nobody taking responsibility.”
• Historically, state‑led technology roll‑outs (e.g., electricity, rail) gave clear lines of responsibility; today, private‑sector‑driven AI blurs those lines.
• Two core imperatives: (1) Companies must be accountable externally, not only to self‑appointed bodies; (2) Citizens must demand accountability and avoid “passing the job onto somebody else.”
• Governments must understand public needs and act as the democratic anchor for AI governance, despite the difficulty of doing so in a fast‑moving technical landscape.

“The formula that has worked for democratic governments for centuries – listening, delivering, and being held to account – must now be applied to AI.”

2. AI as a Growth Engine for the Global South

Amitabh Kant (NITI Aayog) follows, positioning the summit as the first major AI event hosted in the Global South. He highlights three pillars: accessibility, affordability, and accountability, and underlines the multilingual nature of AI as a prerequisite for inclusive impact.

Key insights from Kant

• Data contribution: India supplies ~33 % more training data to large language models (LLMs) than the United States, giving the Global South a lever to influence model behaviour.
• Digital Public Infrastructure (DPI): India’s open‑source, open‑API, interoperable DPI (e.g., UPI, Aadhaar) enabled a seven‑year leap comparable to 50 years of progress elsewhere. The same approach should be replicated for AI – a layer of digital public identity on top of which private innovators can compete.
• Risk of inequality: If AI follows the U.S. trajectory (high GDP share but weak safety nets), the Global South may end up with deeply unequal societies.
• Sectoral impact: AI can transform education, health, nutrition, and climate resilience for the “next billion” by delivering quantum jumps in productivity and service delivery.

“If AI is not multilingual and not affordable for those below the poverty line, we will simply entrench existing inequities.”

3. AI for Climate Action – The Double‑Helix Perspective

Arunabha Ghosh (CEEW) takes the floor to discuss the interdependence of AI and decarbonisation. He frames the relationship as a double helix: the two revolutions must intertwine rather than run parallel.

Core arguments

1. Synergies – AI can optimise renewable integration, improve flood forecasting, and enhance agricultural resilience.
2. Resource footprints – Data‑centres demand energy, water, cooling and land, requiring future‑proof infrastructure that accounts for perpetual water and energy needs.
3. Equity risk – The Global South could become a data‑extraction zone without capturing value.

Proposed guardrails

• Reject the false binary of “climate action vs AI ambition.”
• Embed environmental metrics (energy‑, water‑, lifecycle footprints) into AI strategies.
• Incentivise local, energy‑efficient AI applications and build capacity where impact is maximised.
• Ensure equitable global AI governance, drawing on legal traditions worldwide (reference to Hammurabi’s Code as a precedent for multi‑source rule‑making).

4. African Perspective – Connectivity, Data‑Center Gaps & South‑South Cooperation

A regional representative (identified as “Selma”) from the UN Foundation outlines Africa’s unfinished business:

• Connectivity & energy remain the primary bottlenecks (≈ 600 million Africans lack electricity).
• Data‑center concentration: 1 % of global data‑centers are in Africa; 50 % reside in South Africa alone, leaving many countries unable to host AI workloads.
• Talent pool: Africa’s youthful demographic (median age ≈ 19.7 years) provides 1.2 M GitHub contributors from Nigeria, 300 k from Kenya, etc.
• Access pricing: Kenya pays ≈ US $20 per month for ChatGPT, a prohibitive cost for most.
• Collaboration model: Emphasises regional & South‑South cooperation, leveraging UN‑facilitated partnerships and shared‑value projects rather than duplicated infrastructure spending.

5. UN‑Level Governance, Capacity Building & the Three‑Pillar Approach

Amandeep Singh Gill returns to discuss the UN’s three‑pillared AI strategy (Science, Policy, Capacity).

• Science: Establish an interdisciplinary “AI Science Panel” (akin to the IPCC) that delivers fast, iterative assessments.
• Policy: Align with the Global Digital Compact to create standards, open‑source frameworks, and interoperable APIs.
• Capacity: Build national and sub‑national capacities to govern, adopt, and innovate with AI, avoiding a “K‑shaped” outcome where only a few reap benefits.

He stresses that inclusive dialogue—with all 193 UN member states and intergenerational representation—is essential for legitimacy, speed, and accountability.

6. Responsible AI for Workers – The Value‑Chain View

Safiya Husain (Karya) is asked about responsible AI from the perspective of AI‑industry workers (data annotators, moderators, etc.). She structures her answer around a labour → data → model → evaluation feedback loop.

• Labor rights: Demand fair wages, psychosocial support (especially for moderation tasks), dignity, and upskilling pathways.
• Data inclusivity: Current datasets are Anglophone‑centric; workers in India, Kenya, Uganda, the Philippines are providing labour but not their cultural context.
• Model intentionality: AI must be multilingual, culturally grounded, and evaluated in local languages (e.g., medical chatbots tested with ASHA workers in Bhojpuri).
• Community‑driven tagging: Karya’s Gates‑funded project engaged 20,000 women across six Indian language groups to define gender‑bias categories—revealing that bias manifests differently (community & personality‑based rather than purely occupational).

Safiya calls for ownership across the whole value chain, insisting that workers become co‑creators, not just extractors.

7. Audience Q & A (selected questions)

Questioner	Theme	Respondent(s)	Core Answer
Dinesh Gupta	“AI destructive → constructive”	Panel (mainly Kant)	Emphasised AI as a transformative tool; need guardrails, capacity, and inclusive data to harness it positively.
Rupesh	AI, crypto, dark‑web & cyber‑crime	Kant (with brief comment)	Stressed intersection of technologies (AI + social media) as a vector for abuse; urged holistic, cross‑tech regulation rather than siloed bans.
Yogini (to Ghosh)	Decarbonisation technologies	Ghosh	Highlighted AI‑enabled optimisation of renewables, water‑light cooling for data‑centres; called for resource‑aware AI design.
Istuthi Shavasu	AI for Indic Renaissance	Kant	Advocated leveraging India’s data, talent, and open‑source stack to build indigenous LLMs for cultural preservation and economic growth.
Vandana (senior‑citizen perspective)	AI for older adults	Kant (later)	Shared personal example of using Claude for daily tasks; urged AI literacy programmes for seniors and affordable tooling.
Audience (multiple)	Crypto, dark‑web, cyber‑crime	Kant – brief	Noted the risk of AI tooling being weaponised when paired with other tech; called for coordinated governance.
Audience (AI & youth)	Youth as builders & beneficiaries	Kant & others	Emphasised inclusive participation of young people in policy‑making, South‑South collaborations, and capacity‑building.
Audience (environmental sustainability of AI)	UN‑UNEP coordination	Kant (closing)	Announced a new UN resolution on AI’s environmental footprint; invited partnership with UN‑EP for standards on data‑centre energy & water use.
Audience (frugal AI)	Resource‑constrained AI	Vishal Tripathi (to Ghosh)	Defined Frugal AI as optimised for resources, use‑cases, and choke points; stressed air‑cooled vs water‑cooled trade‑offs and the release of a report on India’s data‑centre ecosystem.

8. Closing Reflections

• Kant reiterates that India’s “AI mission” should blend massive data, talent, and open‑source software to avoid the “new colonisation” of foreign‑owned models. He cites the 11 GW / 5 million‑GPU compute commitment, but warns that software ingenuity will be the decisive factor.
• Gill stresses the need for intergenerational design – the coming AI‑literate youth must steer the trajectory, avoiding a K‑shaped economy.
• Tripathi (CEEW) re‑highlights the Frugal AI framing from the Economic Survey, positioning it as a sustainability imperative rather than merely a constraint. He announces the forthcoming CEEW‑Systemic report on scaling India’s data‑centre ecosystem.

The session ends with thanks to all speakers, the UN Foundation, and the audience, followed by a brief recognition of the USG’s (United Nations) participation.

Key Takeaways

• Accountability must be externalised: Companies need independent oversight; citizens must be empowered to demand responsibility; governments act as the democratic anchor for AI governance.
• Data sovereignty is a lever: India (and the Global South) contributes ~33 % more training data than the U.S.; leveraging this to train locally‑relevant models can reduce dependence on foreign AI providers.
• Digital Public Infrastructure (DPI) model for AI: Open‑source, interoperable layers (e.g., digital identity) can catalyse private‑sector innovation while ensuring public control.
• AI‑climate synergy vs footprint: AI offers powerful tools for decarbonisation (renewable optimisation, flood forecasting) but its energy‑water footprint demands resource‑aware design and standards.
• Africa’s bottlenecks: Massive gaps in connectivity, electricity, and data‑center capacity; solution lies in regional & South‑South cooperation, not isolated infrastructure builds.
• UN three‑pillar strategy: Science panel → policy standards → capacity building is the roadmap to legitimate, fast, and inclusive AI governance.
• Responsible AI for workers: Fair wages, mental‑health safeguards, skill‑upskilling, and inclusive data are essential to avoid an exploitative AI value chain.
• Frugal AI is optimised AI: design for resource constraints, choose air‑cooled vs water‑cooled solutions wisely, and tailor models to local use cases.
• Youth & intergenerational design: Empower young innovators and include them in governance to prevent a future K‑shaped society where benefits accrue only to a privileged few.
• Policy bridging for emerging techs: Address cross‑technology risks (AI‑crypto‑dark‑web) through holistic, intersection‑focused regulation rather than siloed bans.

These points capture the depth and breadth of the panel’s discussion on how AI can be harnessed responsibly to uplift the “next billion” while safeguarding equity, sustainability, and democratic accountability.

See Also:

• democratizing-ai-resources-and-building-inclusive-ai-solutions-for-india
• ai-for-everyone-empowering-people-businesses-and-society
• ai-innovators-exchange-accelerating-innovation-through-startup-and-industry-synergy
• democratizing-ai-resources-in-india
• inclusion-for-social-empowerment
• ai-for-inclusive-economic-progress-the-public-services-ai-stack
• ai-for-inclusive-societal-development
• ai-and-indias-economic-growth-sectoral-impact-and-the-road-ahead
• building-resilient-sustainable-ai-infrastructure-for-people-planet-and-progress
• responsible-ai-at-scale-governance-integrity-and-cyber-readiness-for-a-changing-world