Science ยท GS3
AIKosh and India's AI-ready data ecosystem
One line
The IndiaAI Mission held a workshop on AIKosh, its national platform for datasets and models, with emphasis on discoverable, interoperable, and secure access.
Summary
AI needs data that people can find, understand, use lawfully, and compare. AIKosh raises the issue of data as public infrastructure, but datasets need provenance, quality checks, consent, access controls, and rules for reuse.
PYQ pattern
UPSC technology and governance questions often test digital public infrastructure through data, rights, interoperability, and institutional capacity. See the official UPSC archive.
Core notes
- Metadata explains what a dataset contains, how it was collected, its units, coverage, limitations, and update cycle.
- Interoperability lets systems exchange and interpret data using common formats and standards.
- Secure access can mean controlled permissions, authentication, logging, and privacy protection.
- A national platform can reduce duplication and help startups and researchers, but openness must be matched with rights and accountability.
Prelims lens
- A dataset is not the same as a database management system or an AI model.
- Interoperability concerns compatible exchange and interpretation.
- Discoverability does not mean unrestricted public disclosure.
MCQ 1
Good dataset metadata can describe:
- Coverage and collection method.
- Units, limitations, and update cycle.
- A. 1 only
- B. 2 only
- C. Both 1 and 2
- D. Neither 1 nor 2
Reveal answerHide answer
C
Metadata helps users understand how a dataset can and cannot be used.
MCQ 2
Which statement is correct?
- A. Discoverable data must always be fully public without access controls.
- B. Interoperability means systems can exchange and interpret information using compatible standards.
- C. An AI model and a dataset are identical.
- D. Privacy is unrelated to public data platforms.
Reveal answerHide answer
B
Interoperability requires compatible exchange and interpretation. Access and privacy rules still apply to sensitive data.
Mains
Question, 10 marks, 150 words: How can India build a public-interest data ecosystem for AI while protecting privacy and data quality?
Approach:
- Explain datasets, metadata, interoperability, and AIKosh.
- Discuss research, startups, public services, and sovereign AI capacity.
- Examine consent, provenance, bias, re-identification, access control, and accountability.
- Recommend standard metadata, privacy-by-design, independent audits, tiered access, and user feedback.