AI for All: India’s Policy Architecture for Public-Interest AI and Inclusive Development

Detailed Summary

• Opening Remarks (Prof. Rekha Saxena & co‑conveners)

  • Emphasised that AI is no longer a “future promise” but a present structural feature of Indian society—affecting public services, markets, risk‑assessment, and individual opportunity.
  • Stressed the central question: What should AI do, and under what institutional, ethical, and policy frameworks?
  • Highlighted the need for inter‑disciplinary dialogue (policy, technology, law, economics, health, gender) to avoid siloed thinking.

• Moderator Introduction (Prof. Sanjeev H.M.)

  • Briefly introduced the panel, outlined the agenda (four thematic “sectors”), and delegated the first speaking slot to Dr. Ankit Rajpal.

2. Technology & Ethics Perspective – Dr. Ankit Rajpal

• Public‑Interest AI Definition

  • Public‑interest AI policies must answer three core questions: who are the beneficiaries, who designs the systems, and how they are implemented.

• Bias & Non‑Neutrality of Technology

  • Quoted scholar Ruha Benjamin: “Technology mirrors the system that creates it.” AI must reflect constitutional values, not existing societal biases.

• Sector‑Specific Illustrations

  • Healthcare – AI should not replicate insurance‑driven treatment decisions; rather, it must support equitable clinical choices.
  • Agriculture – Timely AI‑driven advisories can empower farmers, illustrated through the Bhashini multilingual tool that lowered language barriers in Odisha.

• Explainable & Accountable AI

  • Argues for explainable AI so that citizens understand why a decision is made (e.g., why a claim is rejected).
  • Calls for empathy in AI responses, noting the current gap between human service (banker) and automated replies (AI).

• Key Recommendation

  • Public‑interest AI for “Viksit Bharat 2047” must be explainable, reliable, and accountable, built through inclusive stakeholder engagement.

3. Economic & Policy Perspective – Dr. Animesh Naskar

• Closed vs. Open AI Models

  • Closed AI (e.g., ChatGPT, Gemini) risks concentration of power → heightened inequality.
  • Open AI (open‑source) promotes democratization but raises security concerns.

• AI as a Public Good

  • Highlighted the classic public‑good attributes: non‑rivalry, non‑excludability, and non‑excludability of benefits.

• Digital Divide

  • Emphasised that bridging digital gaps is prerequisite for AI to reduce inequality, especially by addressing asymmetric information in markets.

• Policy Alignment with “Amrit Kal” (the government’s flagship initiative)

  • Deconstructed the nine‑letter acronym, linking it to Research, Innovation, Technology, Knowledge, AI, and Leadership.
  • Noted India’s R&D spending (0.6 % of GDP) lags behind China (2.2 %) and other advanced economies, urging greater research investment.

• Future of Work

  • AI will redesign job roles, not merely replace them; thus, skill‑retraining and labour‑market reforms are essential.

• Recommendation

  • Formulate AI policy that balances openness with security, invests heavily in research, and actively reduces digital exclusion.

4. Gender & Global‑Policy Lens – Ms. Sudeshna Mukherjee (UN Women)

• Systemic Biases in Data & Design

  • Cited Invisible Women (Caroline Pérez Mora) to illustrate how historic male‑centric designs (e.g., vehicle crash‑test dummies, HVAC settings) embed gender bias.

• AI‑Induced Discrimination

  • Warned that AI can amplify existing disparities across gender, caste, disability if unchecked.

• Two Immediate Actions

  1. Mandatory Impact Assessment – Conduct pre‑deployment gender‑impact reviews (not post‑hoc).
  2. Women at the Design Table – Ensure female representation in AI development teams.

• Call to Action

  • Emphasised the importance of inclusive policy‑making during the five‑day conference and beyond, as each target group has distinct needs.

5. Fiscal‑Policy & Finance Perspective – Mr. Sudhir Goenka

• AI as a “Fiscal Bridge”

  • Proposed a five‑pillar framework: Transparency, Inclusivity, Equity, Accountability, Sustainability.

• Revenue Generation & Expenditure Efficiency

  • AI can model policy outcomes, eliminate duplication, and predict implementation lags, thereby improving both sides of the fiscal ledger.

• Case Study: Direct Benefit Transfer (DBT)

  • Integration of AI with the JAM Trinity (Jan Dhan, Aadhaar, Mobile) saved ₹3.48 lakh crore (≈ $4.2 bn) and cut administration cost from 16 % to 9 %.

• GSTN Anomaly Detection
  – AI flagging of risky transactions enhances tax compliance and reduces bogus entities, strengthening fiscal space.

• Policy Implication

  • AI should be positioned not as a revenue‑extraction tool but as a structural enabler of transparent, participatory fiscal governance.

6. Innovation & Governance – Mr. Anirudh Gupta (Deloitte India)

• Current Gaps

  • Identified three critical voids: Data & Infrastructure, Talent & Skills, Regulatory & Governance.

• AI Policy Labs

  • Proposed state‑level and national “AI Policy Labs” comprising policymakers, academia, and industry to conduct social‑impact assessments prior to deployment.

• Data Gatekeeping & Ethics Council

  • Advocated for a National AI Ethics Council and a Citizen Grievance Portal to monitor bias, misuse, and ensure purpose‑aligned data usage.

• Open‑Source Platforms

  • Highlighted Deloitte’s launch of Gen W‑AI, a platform for experimentation, encouraging public‑private collaboration.

• Recommendation

  • Institutionalise sandbox environments where public feedback loops continuously refine AI solutions.

7. Humanities & Philosophy of AI – Mr. Sudip (Hindu College)

• Coupling Humanities & Technology

  • Argued for training humanities students in AI fundamentals and embedding ethical considerations at the algorithmic design stage.

• Rupture with the Past

  • Cited philosopher Henri Bergson: memory underpins human experience; generative AI threatens to dilute lived memory, prompting a need to renegotiate our relationship with the past.

• Assessment Redesign

  • Since AI can generate essays, traditional evaluation must evolve → alternative assessment methods that test deeper reasoning rather than rote production.

• AI Rights Debate

  • Raised provocative question: If machines attain consciousness, what rights should they hold? – a cautionary note on future‑proofing ethics.

• Takeaway

  • Philosophical reflection must accompany technical development to safeguard human dignity.

8. “ABCD of AI for All” – Mr. Udit Goenka

Letter	Pillar & Key Points
A – Agentic AI	Move beyond generative text to autonomous agents that perform end‑to‑end tasks (e.g., booking, calls, negotiations).
B – Builders‑First	Provide incentives, GPU access, and funding for Indian creators to develop responsible AI solutions.
C – Cultural & Linguistic Diversity	AI must support all Indian languages (Hindi, Bengali, Tamil, Telugu, Kannada, Assamese, etc.) so a farmer can negotiate in his mother tongue.
D – Digital Public Infrastructure	Leverage existing DPI (Aadhaar, UPI, JAM) to embed AI responsibly within the nation’s digital backbone.
E – Ecosystem	Foster a co‑ordinated ecosystem of academia, industry, and government to sustain AI innovation.

• Call to Action: Scale local builder communities, adopt the TRI framework (Trust, Adaptability, Inclusion), and ensure AI tools are culturally resonant.

9. Closing Remarks & Synthesis

• Moderator (Prof. Sanjeev H.M.) thanked participants, reiterated that public‑interest AI must be accountable, and highlighted the consensus on three cross‑cutting imperatives: explainability, inclusivity, and governance.

• Departmental Endorsement (Prof. Rekha Saxena) emphasized the role of the University of Delhi’s academic ecosystem in shaping AI policy and delivering on the vision of Viksit Bharat 2047.

• Group Photo & Final Acknowledgement – The session concluded with a call for continued collaboration across sectors.

Key Takeaways

• AI must be framed as a public‑policy instrument, not merely a technological novelty.
• Explainable, accountable AI is essential to build citizen trust, especially in sectors like health, finance, and welfare.
• Closed AI models risk power concentration; an open‑source, secure approach is needed to democratize benefits.
• Gender and other systemic biases are entrenched in data and design; mandatory impact assessments and inclusive design teams are non‑negotiable.
• Fiscal integration of AI (e.g., DBT‑JAM‑AI) can yield massive savings and improve compliance, demonstrating AI’s role as a fiscal bridge.
• Policy Labs, Ethics Councils, and grievance portals should be institutionalised to continuously monitor bias, misuse, and societal impact.
• Humanities and philosophy must inform AI development, prompting new assessment models and pre‑emptive ethical debates (e.g., AI rights).
• Agentic AI, builder‑first incentives, multilingual capability, DPI leverage, and ecosystem coordination constitute the “ABCD” roadmap for inclusive AI in India.
• Research investment must be dramatically increased (target > 2 % of GDP) to keep pace with global competitors and empower AI innovation.
• Cross‑sector collaboration—government, academia, industry, and civil society—is the cornerstone for achieving Viksit Bharat 2047 through public‑interest AI.

---

See Also:

• shaping-secure-ethical-and-accountable-ai-systems-for-a-shared-future
• building-public-interest-ai-catalytic-funding-for-equitable-access-to-compute-resources
• ai-for-everyone-empowering-people-businesses-and-society
• ai-impact-forum-democratising-ai-resources
• ai-innovators-exchange-accelerating-innovation-through-startup-and-industry-synergy
• enterprise-adoption-of-responsible-ai-challenges-frameworks-and-solutions
• founders-funders-the-india-ai-capital-ecosystem
• ai-for-fraud-prevention-and-financial-inclusion-in-bfsi
• navigating-the-ai-regulatory-landscape-a-cross-compliance-framework-for-safety-and-governance
• policymakers-dialogue-on-ai-policy-evolution-and-the-rule-of-law