India's AI policy in 2026 centres on the M.A.N.A.V. framework announced by Prime Minister Narendra Modi at the India AI Impact Summit 2026 (16-20 February 2026, New Delhi), the IndiaAI Mission (approved March 2024), the Digital Personal Data Protection Act 2023 (DPDP Act), and MeitY advisories on generative AI.
M.A.N.A.V. is India's values-based AI governance framework introduced by PM Narendra Modi at the India AI Impact Summit 2026. It complements the G20 AI Principles (New Delhi Leaders' Declaration, 2023) and UNESCO's Recommendation on the Ethics of AI (2021).
| Pillar | Meaning |
|---|---|
| M | Moral and Ethical Systems — fairness, transparency, human oversight |
| A | Accountable Governance — transparent rules, institutional oversight |
| N | National Sovereignty — data security, domestic compute, indigenous AI |
| A | Accessible and Inclusive AI — broad-based multiplier, not elite monopoly |
| V | Valid and Legitimate Systems — verifiable, lawful, transparent deployment |
Reference: Press Information Bureau, Government of India, India AI Impact Summit 2026.
| Instrument | Authority | Status |
|---|---|---|
| IndiaAI Mission | MeitY, INR 10,371.92 crore approved | Active (March 2024-2029) |
| DPDP Act 2023 | MeitY, Data Protection Board | Act passed; Rules notified 2025 |
| MeitY Advisory on GenAI (March 2024) | MeitY | Non-binding guidance |
| M.A.N.A.V. Framework | Government of India | Announced Feb 2026 |
| Digital India Act (draft) | MeitY | Under consultation |
| NITI Aayog Responsible AI Principles | NITI Aayog | Reference (2021) |
| RBI Guidance on AI in Banking | Reserve Bank of India | 2024 circular |
| Violation | Max Penalty |
|---|---|
| Failure to notify personal data breach | INR 200 crore |
| Failure to protect children's data | INR 200 crore |
| Failure of Significant Data Fiduciary duties | INR 150 crore |
| Any other violation | INR 50 crore |
| Overall cap per instance | INR 250 crore |
IndiaAI Compute Portal — Launched January 2025, offering subsidised GPU access (34,000+ GPUs procured) to startups, researchers, and PSUs under the IndiaAI Mission.
BharatGPT, Krutrim, Sarvam AI, Ola Krutrim — Indigenous foundation models supported under the IndiaAI Mission's "Make AI in India" pillar.
MeitY Advisory (15 March 2024) — Asked "significant" platforms to seek explicit government permission before deploying under-testing generative AI models; the advisory was updated on 15 March 2024 to exempt startups after industry feedback.
Every AI product targeting Indian users in 2026 must:
Q: Is M.A.N.A.V. a law? No — it is a values framework. It guides regulation and is cited in MeitY policy documents.
Q: When does the DPDP Act come into force? The Act was notified 11 August 2023; the Rules were notified in January 2025 with a phased implementation.
Q: What is a Significant Data Fiduciary? A class of data fiduciary notified by the government based on volume, sensitivity, and impact.
Q: Does India require localization of AI training data? The DPDP Act allows cross-border transfer except to countries the Central Government blacklists.
Q: Is there an India equivalent of the EU AI Act? Not yet — the Digital India Act is expected to include AI provisions in 2026-2027.
Q: Does IndiaAI funding require M.A.N.A.V. alignment? Projects funded under the IndiaAI Mission must demonstrate alignment with responsible AI principles.
Q: What about deepfakes in India? Section 66E (IT Act) plus IPC/BNS provisions apply; MeitY's November 2023 advisory requires due diligence.
India's AI policy in 2026 is values-first (M.A.N.A.V.), mission-driven (IndiaAI), and anchored by the DPDP Act. Companies that build transparency, explainability, and data sovereignty into their products from day one will ship faster across India's USD 17 billion AI market.
Build India-ready AI with Misar AI's M.A.N.A.V.-aligned governance templates.
Free newsletter
Join thousands of creators and builders. One email a week — practical AI tips, platform updates, and curated reads.
No spam · Unsubscribe anytime
A practical 2026 AI ethics checklist covering governance, data, model, deployment, and monitoring — aligned with NIST, I…
A practical 2026 guide to detecting and mitigating AI bias: statistical tests, fairness metrics, tooling (AIF360, Fairle…
Comments
Sign in to join the conversation
No comments yet. Be the first to share your thoughts!