Democratisation of AI – Bharat Story

Detailed Summary

  • The moderator welcomed the panel, highlighted the India AI Impact Summit and stressed the need to discuss “democratisation of AI” within the context of the Bharat Stack.
  • Cited macro‑level statistics: 14 billion UPI transactions, ≈1.4 bn citizens across ≈700 districts and 22 languages – the world’s largest digital user base.
  • Posed three guiding questions: (a) What is AI doing for the “farm‑to‑finance” pipeline? (b) What partnerships and platforms are required? (c) How can entrepreneurs, think‑tanks and students plug into these efforts?

2. Madhya Pradesh: AI in Agriculture & Welfare (Speaker: Mr. Sanjay Dubey / “Anshuman”)

2.1 Satellite‑based Crop Intelligence (“fourth cycle”)

  • Traditional Girdawari (field‑survey by Patwari) replaced by ML‑driven satellite classification with ≈95 % accuracy.
  • Automatic generation of crop‑type, area, and expected yield reduces omission & commission errors.

2.2 Yield & Disaster‑Response Gains

  • Example: Soybean sowing adjustments based on climate‑shift predictions produced a ≈25 % yield lift over three years.
  • Hail‑storm case study: Manual compensation estimate ₹18.5 cr vs. satellite‑based estimate ₹4.5 cr – no farmer complaints, demonstrating trust in AI‑enabled verification.

2.3 Direct‑Benefit‑Transfer (DBT) & Zero‑Leakage Ambitions

  • Discussed the Samagra ID system (unique, family‑linked identifier since 2012) that underpins ≈400 state schemes.
  • Highlighted plans to open anonymised land‑records (Sampada), property transactions, and digitised land‑revenue data to startups for AI‑driven analytics (e.g., price‑prediction, fraud detection).

2.4 GIS‑Based Service Delivery

  • Piloted a routing engine for delivering milk to ≈80 k Anganwadis; the system computes optimal routes in ≈1 hour, demonstrating AI‑enabled logistics at scale.

3. Odisha: Language, Localization & Citizen‑Centric Platforms (Speaker: Mr. Vishal Kumar Dev)

3.1 Language‑Data Infrastructure

  • Odisha’s low‑resource language (Odia) – only a few datasets available.
  • Initiative: Conversion of 1,600 historical texts into digital form, contributed to OI Kosh (India AI Mission).
  • Upcoming “Odia Bhashadhan” program to crowd‑source data across 13 sectors (tourism, history, agriculture, etc.) from experts, villagers, and scholars.

3.2 Voice‑Dataset for Dialects

  • Created district‑wise voice corpora (e.g., Santali variations) to improve speech‑AI for regional dialects.

3.3 Sarvam Platform – Rice‑Fallow Management

  • Hyper‑local soil & weather mapping for ≈5 million calls to farmers, delivering real‑time advisories and encouraging multi‑cropping.

3.4 Skilling & Outreach

  • AI for All – a free, 4.5‑hour course for the general public.
  • University collaborations (e.g., Odisha University of Agriculture & Technology) to produce sector‑specific AI datasets (agri, health, disaster‑management).
  • Capacity‑building partnership with the Vadhwani Foundation to train government officers from clerical to secretary level.
  • Recent rollout of AI basics to 12,000 schools for high‑school students.

4. Karnataka: Skilling, Centres of Excellence & DPI‑Enabled Innovation (Speaker: Dr. Manjula N)

4.1 “Nipuna” – Industry‑Aligned Upskilling

  • Public‑private model: 40 % of training cost funded by the government, rest by industry partners.
  • Focus on job‑ready AI, VLSI, and data‑science skills for fresh graduates.

4.2 Centres of Excellence (CoE)

  • 20+ CoEs spanning AI, biotech, cybersecurity, and applied AI.
  • CoEs act as research hubs, innovation incubators, and industry‑collaboration nodes. Examples: AI CoE at IIT IIIT Dharwad, Biotech AI CoE under CCAMP.

4.3 Academic‑Industry Linkages

  • Partnerships with RV University, Manipal University, Alliance University for quantum‑computing labs and AI curriculum development.

4.4 DPI Assets & “E‑Sahamati” Portal

  • Karnataka’s E‑Governance department curates massive datasets (farmer‑produce, health, PDS, forest, smart‑city).
  • E‑Sahamati – a one‑stop portal where startups can request data access under governance, security, and privacy safeguards.

4.5 Open‑Source Identity Platform (MOSIP)

  • Development of MOSIP (Modular Open‑Source Identity Platform) at IIIT Bangalore, now adopted by ≈35 countries.

5. Telangana: Policy, Ecosystem & Zero‑Cost Access (Speaker: Mr. Fani Nagarjuna)

5.1 AI Policy 2024

  • Comprehensive AI policy covering responsible, inclusive AI, and mandating zero‑cost access to foundational models.

5.2 Ecosystem Building

  • ₹1,000 cr “Startup Fund‑of‑Funds” plus private co‑investment.
  • Co‑innovation programmes linking industry, academia, and startups.

5.3 Cloud & Compute Infrastructure

  • Partnership with AWS for a $67 bn investment in cloud & data‑center infrastructure over the next decade.

5.4 Structural Recommendation

  • AI initiatives should be housed in an autonomous, agile entity (e.g., “I Come at Davos”) that can scale like a startup while retaining governmental backing.

5.5 Data‑Exchange (TGDEX)

  • Launched TGDEX – Telangana Data Exchange, a sandbox with >1,100 datasets and pre‑built ML models, delivering zero‑cost AI tooling to startups.

6. Cross‑State Collaboration & Startup Showcase

StateNotable Startup / Use‑Case Mentioned
KarnatakaCoreover.ai – AI chatbot (Bharat GPT) for IRCTC, NPCI, Indian Navy; overcomes language barriers.
KarnatakaServerMai – Voice‑assistant for Aadhaar‑based grievance redressal; multilingual Indic‑GenAI.
OdishaRice‑Value‑Management (Sarvam) – AI‑driven call‑center for crop advisory (5 M calls).
Madhya PradeshGIS routing engine for Anganwadi milk delivery (80 k points).
KarnatakaRemedio – AI solution for diabetic‑retinopathy detection.
KarnatakaCure Bay – Healthcare AI platform.
KarnatakaODIA‑Reading‑Fluency – AI tool for early‑grade literacy.

7. Q&A Highlights

  1. Zero‑Leakage Welfare – Odisha’s SUBHDRA scheme (₹50 k per woman, ~1.14 cr beneficiaries) and historic ghost‑ration‑card purge (≈2 lac eliminated) illustrate AI‑enabled verification.

  2. Data‑Sharing Philosophy – Both Madhya Pradesh (Samagra ID, Sampada) and Karnataka (E‑Sahamati) emphasised anonymised, open data as the foundation for AI startups.

  3. Skill Gap in Karnataka – Despite ≈2.5 mn technically qualified residents, the state is addressing skill‑gap via Nipuna, CoEs, and AI‑for‑All programmes.

  4. MSME Impact – Panel urged focus on AI for micro‑enterprises: a 15‑18 % CAGR boost could lift the sector’s contribution to ≈$4.5 trn by 2030. Karnataka’s Gen AI programme pairs students, industry, and AI mentors for rapid prototyping.

8. Closing Statements (One‑line Vision per State)

  • Karnataka (Dr. Manjula N) – “Human capital is our biggest asset; skilling, up‑skilling and AI‑enabled solutions will drive societal benefit.”
  • Odisha (Vishal Dev) – “From a mineral‑driven economy to a mind‑driven AI economy—Odisha will become the world’s AI proving ground.”
  • Madhya Pradesh (Sanjay Dubey) – “Precision in delivery, dignity in service access.”
  • Telangana (Fani) – “Zero‑cost AI access through robust policy, data‑exchange and compute infrastructure.”

The session concluded with a brief thank‑you, a group photo, and applause.

Key Takeaways

  • AI democratization hinges on three pillars: (1) Policy & governance, (2) Open, high‑quality data (DPI), (3) Affordable compute & skill development.
  • Satellite‑derived crop intelligence in Madhya Pradesh now achieves ≈95 % classification accuracy, yields a ≈25 % increase and cuts disaster compensation by ≈75 %.
  • Odisha’s language‑first strategy (1,600 digitised texts, district‑wise voice corpora, Odia Bhashadhan) tackles the low‑resource language barrier essential for inclusive AI.
  • Karnataka’s “Nipuna” and 20+ Centres of Excellence illustrate a public‑private upskilling model that subsidises 40 % of training costs and aligns curricula with industry demand.
  • Telangana’s AI policy (2024) and TGDEX sandbox provide zero‑cost AI model access to startups, supported by a $67 bn AWS cloud investment.
  • Data‑exchange portals (E‑Sahamati, TGDEX, Samagra ID) enable anonymised sharing of massive public datasets (land records, health, PDS, agriculture) for AI innovation.
  • Cross‑state collaboration is already materialising: startups such as Coreover.ai and ServerMai leverage Bharat Stack APIs to deliver multilingual, AI‑driven citizen services.
  • MSMEs represent a massive AI opportunity; a 15‑18 % productivity lift could add ≈$4.5 trn to India’s GDP by 2030.
  • Trust in AI grows when transparent, satellite‑based verification replaces manual bureaucratic processes, as evidenced by Madhya Pradesh’s hail‑storm compensation case.
  • Future roadmap: build an autonomous AI‑innovation agency (as advocated by Telangana) that can scale across states, ensure responsible AI, and keep the cost of access effectively zero.

Note: Mr. Fani (Telangana) and Mr. Bhaskar Katamneni (Andhra Pradesh) were referenced but only Fani participated in the discussion. Their presence is flagged for completeness.

See Also: