India AI Impact Summit 2026

ai-capacity-building-scaling-knowledge-driving-innovation

6 min read

TRACK:Human Capital
SPEAKERS
  • Dr. Mohammed Misbahuddin, C-DAC Bangalore
  • Dr. S D Sudarsan, C-DAC Bangalore
  • Mr. Ramesh Naidu Laveti, C-DAC Bangalore
  • Mr. Srikanth Chandrasekaran, IEEE
SESSION OVERVIEW

As AI rapidly reshapes education and research, academic institutions face growing challenges in integrating AI literacy, advanced skills, and ethical readiness into curricula and research programs. This workshop addresses these challenges by bringing together academic leaders, researchers, and standards bodies to explore scalable, interoperable frameworks for AI capacity building.

VIDEO RECORDING

AI Capacity Building - Scaling Knowledge, Driving Innovation

Abstract

The workshop explored how academic institutions, standards bodies and research organisations can cooperatively build scalable, interoperable AI capacity. Panelists described IEEE’s ethical‑design standards, EU–India partnership models, CDAC’s super‑computing infrastructure, and education‑focused initiatives such as AI Yuvakula for school children. The discussion moved through concrete programmes (IEEE 3119/319 standards, AI certified tracks, Grand Challenge on misinformation), hardware‑scale‑up plans (30 + supercomputers, 90 + PFLOPS target, “AI‑in‑a‑box”), curriculum integration of ethics and compliance, and the challenges of bringing AI to domains such as weather forecasting. The session concluded with a brief audience Q&A and a summary of next steps for collaborative capacity‑building.

Detailed Summary

  • Moderator (Ramesh Naidu Laveti) welcomed the panel and introduced the participants.
  • He highlighted the need for a “third way” that blends inclusivity, sovereignty and ethical alignment in AI capacity building.

2. IEEE’s Holistic Approach to AI Skilling

2.1 Ethical‑Design Standards (IEEE 7000 Series)

  • Srikanth Chandrasekaran explained that IEEE has been developing socio‑technical standards since 2016, culminating in the IEEE 7000 series that embed ethical values into system‑level design.
  • The standards address unintended use cases, governance, and accountability.

2.2 Training & Certification Programmes

  • IEEE offers a layered certification pathway (introductory → advanced) that trains individuals to become IEEE‑certified assessors capable of evaluating AI products for data‑privacy, transparency, bias, and accountability.
  • IEEE 3119 (AI procurement) and IEEE 319 (AI procurement standards) were mentioned as tools for government and industry buyers.

2.3 Children‑Focused Initiative – AI Yuvakula

  • In partnership with Meher Care Foundation and Sheikshana Foundation, IEEE launched AI Yuvakula to help school‑age children (≈ 12 years, grade 7) transition from regional languages to English‑medium instruction while learning AI concepts.
  • The program also touches on special‑needs education (e.g., autism) via the e‑Sadhya project.

2.4 Grand Challenge on Misinformation

  • IEEE Standards Association, together with OECD, AI Commons, UNESCO, opened a Grand Challenge (proposals accepted until end‑March).
  • The focus is on combating AI‑driven misinformation/disinformation.
  • CDAC super‑computers will host the training and evaluation phases; CDAC has pledged mentorship to teams needing help porting workloads to high‑performance hardware.

3. Europe–India Collaboration

3.1 Leveraging EU Training Modules

  • Srikanth noted that the European Commission and EIT Campus provide ready‑made AI training modules that can be repurposed for Indian audiences.

3.2 Joint Research & Regulatory Dialogue

  • Dr. Lalit Patil highlighted the EU AI Act (risk‑based classification) versus India’s currently absent AI‑specific law.
  • He advocated a coordinated framework where European risk‑based standards inform Indian compliance models, impact‑assessment practices, and sector‑specific deployment strategies.

3.3 Preventing Talent Drain & Enabling Start‑up Mobility

  • The panel discussed a programme for Indian start‑ups to scale in Europe, offering visas, office space, HPC access and seed funding (up to €1 million).
  • The intent is to nurture reciprocal research corridors while avoiding permanent brain‑drain.

4. CDAC’s Super‑Computing & “AI‑in‑a‑Box”

4.1 Current Infrastructure

  • Gokul Thivdan (CDAC) reported 30 + supercomputers deployed nationwide over the last ~ 7 years, totalling ≈ 60 PFLOPS (with a 90 PFLOPS target by Sep 2026).
  • Notable systems: Irawath (baseline) and a 30 PFLOPS system slated for Bangalore (SPCA).

4.2 “AI in a Box” (Paramshavak)

  • Paramshavak is a tabletop AI workstation that combines GPUs with a complete software stack, aimed at Tier‑2/3 institutions.
  • It allows local researchers and students to run AI workloads without needing a full data‑center.

4.3 Convergence of HPC & AI

  • CDAC is integrating AI optimisation into traditional HPC to improve utilisation efficiency.
  • The National Semiconductor Mission (NSM) is developing indigenous AI chips (DPUs, TPUs, neuromorphic, in‑memory) and planning exascale systems built on these custom silicon.

4.4 Mobile‑First AI Infrastructure

  • Gokul argued that mobile devices will become the dominant AI compute endpoint, with efficiency expected to rise from the current 20‑30 % to near‑full utilisation as models and data quality improve.
  • A grid of AI‑enabled mobile devices could help bridge the rural‑digital divide.

5. Curriculum & Education Integration

  • Dr. Lalit emphasised that EU regulations mandate ethical‑AI training for all employees, suggesting a template for Indian curricula.
  • The panel referenced India’s National Education Policy 2020, which now recognises credit‑based learning from industry‑aligned courses, opening the door for standards‑driven AI modules.

6. AI Procurement Challenges

  • Srikanth revisited IEEE 3119, pointing out that AI‑enabled procurement is now pervasive in government contracts.
  • Capacity‑building is needed so that procurement officers understand the standards and can assess AI solutions for bias, privacy and accountability.

7. Building a Shared AI Literacy Framework

  • The panel distilled three universal pillars for AI competency at the K‑12 level:

    1. Ethics & Humanity (human‑centred AI, inspired by early literature such as Čapek’s R.U.R.)
    2. Sustainability (energy‑efficient model design; note that large‑language‑model training could otherwise require tens of nuclear plants)
    3. Critical Thinking (recognising hallucinations, bias, and the limits of AI)

  • These pillars can underpin a common standard for AI teaching across continents.

8. Industry Adoption of European Regulatory Expectations

  • Dr. Lalit advised that companies should embed data‑governance and traceability from design‑time, not as an after‑thought.
  • Automated compliance tooling and continuous risk‑assessment are key to meeting both EU and Indian expectations.

9. Weather‑Forecasting & AI – Audience Q&A

QuestionSpeaker(s)Summary of Response
Hybrid CPU‑GPU architecture for NCMRWF modelsDr. Anita (audience) – answered by Gokul (CDAC)GPU excels at low‑precision workloads; existing CPU‑centric models need code refactoring and hotspot identification before porting.
Practical experience moving legacy weather models to GPUAudience expert (former forecaster)Porting is non‑trivial; requires module‑by‑module analysis, years of effort for large code‑bases, and validation of AI‑driven components for stability in extreme events.
EU support for climate‑change modellingAudience (EU perspective)EU’s Copernicus data and funding mechanisms (e.g., Horizon Europe) can support joint climate‑AI projects; collaboration via ECMOWF and UK‑MAT is possible.

10. Closing Summary & Forward‑Looking Statements

  • Srikanth gave a concise wrap‑up (≈ 1 minute): IEEE’s global standards are being contextualised for India and Europe, creating a scalable, open platform for capacity building.
  • The panel thanked the audience, reaffirmed commitment to joint projects, and invited continued dialogue.

Key Takeaways

  • IEEE’s standards ecosystem (7000 series, 3119, 319) is central to AI capacity building, offering both ethical guidance and concrete procurement guidance.
  • AI Yuvakula targets multilingual Indian school children, bridging language transition and AI literacy, while e‑Sadhya addresses autism‑focused AI training.
  • The IEEE Grand Challenge (misinformation focus) invites proposals until end‑March; CDAC will provide super‑computing resources and mentorship.
  • CDAC now operates 30 + supercomputers (≈ 60 PFLOPS) and is expanding to 90 PFLOPS; the Paramshavak “AI‑in‑a‑box” makes AI compute accessible to Tier‑2/3 institutes.
  • EU‑India partnership can repurpose European AI training modules, align risk‑based regulatory frameworks, and create reciprocal start‑up pathways with visa, funding and HPC support.
  • Curriculum reform is underway under India’s NEP 2020; ethical‑AI training (as mandated by the EU AI Act) should be embedded from the outset of AI courses.
  • Mobile‑first AI is seen as the future vector for rural inclusion; model efficiency gains will lower compute requirements, enabling large‑scale device grids.
  • AI procurement is a growing governance challenge; capacity‑building around IEEE 3119 is essential for government and industry buyers.
  • Shared K‑12 AI literacy pillars (ethics, sustainability, critical thinking) provide a common foundation for cross‑regional standards.
  • Weather‑forecasting integration requires selective GPU porting, extensive code refactoring, and validation of AI‑augmented components—collaboration between CDAC, NCMRWF and EU data programmes is envisaged.
  • The session concluded with a consensus that open, standards‑driven collaboration is the most effective route to scalable, inclusive AI capacity building across India, Europe and the broader global community.

See Also:

Graph View

Backlinks