PIB TrackerIndia's AI problem is not compute, it is data that can be shared
The IndiaAI Mission's AIKosh workshop put UIDAI, MoSPI, the World Bank and Google DeepMind in one room on the question of which public datasets can lawfully be opened.
What happened
- AIKosh is the national datasets and models platform under the IndiaAI Mission.
- The workshop was inaugurated by the Secretary, MoSPI.
- The stated challenge is data that is discoverable, interoperable and securely accessible.
- Sessions covered lawful data sharing and privacy-preserving use.
- Participants included Google DeepMind, UIDAI, the World Bank and IIT Jodhpur.
For Prelims
- AIKosh: the national datasets and models platform of the IndiaAI Mission.
- IndiaAI Mission: under the Ministry of Electronics & IT, with pillars including compute capacity, datasets, application development and safe and trusted AI.
- Sovereign AI: the framing that a country's data assets and model capability are strategic national resources.
- The bottleneck named: data being discoverable, understandable, interoperable and securely accessible across ministries — governance, not volume.
- NeGD: the National e-Governance Division, which spoke on unlocking data for AI in governance.
- MoSPI: the Ministry of Statistics and Programme Implementation, which inaugurated the workshop.
- Foundational models: large models trained on broad data and adapted to many tasks — the second session's subject.
- Venue: the India Habitat Centre, New Delhi.
For UPSC: The governance layer of the AI debate, which is where India's constraint actually sits. Use it for the IndiaAI Mission, data governance and privacy, digital public goods and sovereign AI.
What it is NOT: A workshop is a consultation, not a data-sharing policy — no category of government data was opened by this event.
For Mains
Syllabus: GS3.13 · GS2.15 · Linkage L2
Anchor
India's AI constraint is not the absence of data but the absence of a lawful, interoperable way to move it between the ministries that already hold it.
Substantiation (data)
The MoSPI Secretary framed the challenge as data being discoverable, understandable, interoperable and securely accessible across all ministries — with AIKosh as the platform.
Exemplification
Placing UIDAI, MoSPI, the World Bank and Google DeepMind on the same panel on lawful sharing shows the question is jointly legal, statistical and technical.
Problematisation
Opening government datasets collides with the Digital Personal Data Protection framework and with the re-identification risk in granular administrative data.
Way-forward
Publish a tiered data classification with purpose-limited access, so sharing is a default with conditions rather than a case-by-case negotiation.
Position
Sovereign AI will be decided in data governance rules, not in GPU counts.
Deploys into: IT and emerging tech + e-governance (GS3.13, GS2.15) · IndiaAI Mission, data governance, digital public goods and privacy.
Ministry of Electronics & IT · 2026-08-20 · PRID 2301666 · PIB source ↗