Every AI Policy Ever · open in the interactive register

Published (non-binding, current)Non-binding · Institutional mandate

IndiaAI Safety Institute (establishment announcement)

Record IN-01

Issued by
Ministry of Electronics and Information Technology (MeitY), Government of India
Jurisdiction
India
Level · region
National · South & Southeast Asia
Published / adopted
2025-01-30
Language
English

Summary

Minister Ashwini Vaishnaw announced India's AI Safety Institute under the 'Safe and Trusted AI' pillar of the IndiaAI Mission. It will address AI risks and safety challenges and pursue research on Indian datasets. It uses a hub-and-spoke model, with academic and research institutions as spokes and private partners joining the hub. Its related calls cover watermarking and labelling, AI risk assessment, stress-testing tools and deepfake detection.

Safety topics

safety instituteevaluation & testingdeepfakes & synthetic mediaresearch priorities

Open question

checker: All stored evidence is an announcement, labelled alt-official, so level B is right. The one open point is date_published: the day is real but its file belongs to another run and is not linked to this record. Attach the PIB release to the record (extra_urls) so the date has accepted evidence Issue #181

The record shows its last checked values until a maintainer decides; the decision is recorded in its history.

Status notes

The Minister of Electronics and IT announced the institute at a media interaction at Electronics Niketan, New Delhi, on 30 January 2025, adopting what he called a techno-legal approach, together with the empanelment of over 18,000 GPUs and a call for proposals for indigenous foundation models; the IndiaAI article on it is dated 31 January 2025. The institute is incubated by the IndiaAI Mission under its Safe and Trusted pillar. The India AI Governance Guidelines of November 2025 (IN-02) describe the recently established institute as the main body for guiding the safe and trusted development and use of AI in India: it should do research, risk assessment and capacity-building, test and evaluate AI systems for risks, advise policymakers and industry, and anchor India's participation in the International Network of AI Safety Institutes.

Sources

Open the official source https://indiaai.gov.in/article/india-takes-the-lead-establishing-the-indiaai-safety-institute-for-responsible-ai-innovation

Secondary source: https://psa.gov.in/CMS/web/sites/default/files/publication/India%20Takes%20the%20Lead%20Establishing%20the%20IndiaAI%20Safety%20Institute%20for%20Responsible%20AI%20Innovation.pdf

Local copy

Partial (secondary source only). The official source could not be downloaded (The site's TLS certificate could not be verified. robots.txt could not be checked, which RFC 9309 treats as 'do not crawl'); only the secondary source was saved. The secondary source can no longer be downloaded (DNS lookup failed, so the site could not be reached from this machine (try another network, or save the document by hand)); the earlier copy is kept. Last attempt 2026-10-18 06:48:02 UTC

Secondary source (third-party copy, not published) archive/IN-01/2026-10-18/secondary.pdf
Retrieved 2026-10-18 02:59:11 UTC
SHA-256 e3a5200d88d0280abbb8333b2e0278c3cadb3b80117c8fc2100c4c9716045602

Verification

B
Another official source, such as an announcement by the issuer, confirms the title, date and status. checked 2026-10-18 · alternative-official-url-fetched
Refresh of 18 Oct 2026. Every source of this record is an announcement about the institute: the IndiaAI article (the official_url), which could not be stored on any attempt (robots.txt unreachable, DNS failures, then a TLS certificate that could not be verified) and of which the Office of the Principal Scientific Adviser's PDF copy (psa.gov.in, the secondary_url, dated 31 Jan 2025) is stored; and the PIB release of MeitY of 30 Jan 2025 (PRID 2097709, pib.gov.in), stored on 18 Oct 2026 in run 2026-10-18-recheck-south-asia-1 but not reachable again in this refresh (DNS failure). No document that is itself the instrument, such as an order establishing the institute, is known, so the record is level B. The day of the announcement (30 Jan 2025) rests on the PIB release. The PSA copy confirms the announcement by the minister, the Safe and Trusted pillar, the mandate, the hub-and-spoke model and the second Expression of Interest themes (watermarking and labelling, AI risk assessment and management, stress-testing tools, deepfake detection) used in the summary. The India AI Governance Guidelines (IN-02) confirm the institute's later role. No stored source gives the institute's legal form or a date of formal operational launch.
Second check: Independent check by Claude Code (claude-sonnet-5-5), 2026-10-18: 3 field(s) supported.

History

  1. 2026-10-18verification.level: A → B — level computed from the evidence by tools/pipeline.py2026-10-18-refresh-level-rule-1 · Claude Code (claude-opus-5-5)
  2. 2026-10-18notes: Announced at a media interaction on 30 Jan 2025 (NeGD record). The PSA/IndiaAI w → The Minister of Electronics and IT announced the institute at a media interactio — rewritten from the stored PIB release, the PSA copy of the IndiaAI article and the guidelines: the media interaction of 30 Jan 2025 and what was announced with it, the techno-legal approach, the institute's role in the November 2025 guidelines; the remark that the legal form and launch date 'were not confirmed' moved to verification.note2026-10-18-recheck-south-asia-1 · Claude Code (claude-opus-5-5)

Related records

India AI Governance Guidelines India · Published (non-binding, current)

International Network of AI Safety Institutes International Network of AI Safety Institutes · Published (non-binding, current)

Citation

Ministry of Electronics and Information Technology (MeitY), Government of India. IndiaAI Safety Institute (establishment announcement). 2025-01-30. https://indiaai.gov.in/article/india-takes-the-lead-establishing-the-indiaai-safety-institute-for-responsible-ai-innovation. Status: Published (non-binding, current); verified 2026-10-18 (level B). Every AI Policy Ever, version 1.18 (snapshot 2026-10-18), record IN-01.

Improve this record

Know of an amendment, a status change or an error? Report news or an error on GitHub; the form opens with this record filled in.