Quick answer
India has no standalone AI law in 2026. AI is governed through existing instruments: MeitY's India AI Governance Guidelines (November 2025), the DPDP Act and Rules for personal data, IT Rules amendments on synthetically generated information notified in February 2026, CERT-In's 2022 incident and logging directions, RBI's FREE-AI framework for regulated entities, and SEBI's AI/ML reporting and accountability rules for market participants.
- MeitY's guidelines favour principles, existing laws and targeted amendments over a new AI statute for now.
- The DPDP Act and Rules govern personal data in AI systems, with core rules applying from 13 May 2027.
- IT Rules amendments on synthetically generated information were notified in February 2026; check the final text for specifics.
- CERT-In's 6-hour incident reporting and 180-day log retention apply to AI infrastructure like any ICT system.
- RBI FREE-AI recommendations bind regulated entities only once RBI issues directions; SEBI already holds users of AI responsible.
- Builders should map each AI system to the sector, data and content rules that apply to it.
Who this is for: AI startup founders, CTOs and ML platform leads, Compliance teams at banks, NBFCs and market intermediaries, Product and trust and safety leads at platforms, Foreign AI companies serving Indian users
How does India regulate AI in 2026?
India does not have an AI Act. Instead, AI is governed by a layered set of instruments, each applying to a part of the AI stack or to a sector.
| Layer | Instrument | What it governs |
|---|---|---|
| Horizontal policy | MeitY India AI Governance Guidelines (Nov 2025) | Principles, institutions and recommendations; not binding law |
| Personal data | DPDP Act, 2023 and DPDP Rules, 2025 | Training data, prompts, outputs and logs containing personal data |
| Online content | IT Act and IT (Intermediary Guidelines and Digital Media Ethics Code) Rules, 2021, as amended | Intermediary due diligence, synthetic content labelling |
| Cyber security | CERT-In Directions of 28 April 2022 | Incident reporting, logs, clock sync for ICT systems including AI |
| Banking and payments | RBI, including the FREE-AI framework report | AI use by regulated entities; payment data storage |
| Securities markets | SEBI circulars and regulations | AI/ML reporting and accountability of regulated persons |
The approach was set out in MeitY's guidelines: rely on existing laws such as the IT Act, DPDP, consumer protection and sectoral rules, and amend them where gaps appear, rather than enacting a new horizontal statute now.
This page summarises each layer and links to detailed guides. For a builder, the practical task is a mapping exercise: for each AI system, which data, content and sector rules apply?
MeitY's India AI Governance Guidelines
Released in November 2025 under the IndiaAI Mission, the guidelines are principle-based and recommend no standalone AI law for now.
Seven sutras
- Trust is the foundation
- People first
- Innovation over restraint
- Fairness and equity
- Accountability
- Understandable by design
- Safety, resilience and sustainability
Six pillars of recommendations
Infrastructure, capacity building, policy and regulation, risk mitigation, accountability, and institutions.
Proposed institutions
- AI Governance Group (AIGG)
- Technology & Policy Expert Committee (TPEC)
- AI Safety Institute (AISI)
Tools it encourages
Voluntary frameworks, techno-legal approaches, grievance mechanisms, an AI incidents database, and graded liability based on function and risk.
The guidelines do not create penalties on their own. Their significance is directional: they tell builders which principles future amendments and sector rules are likely to follow. Full breakdown: India AI Governance Guidelines explained.
The DPDP Act and Rules: the binding layer for personal data
For most AI systems, the DPDP Act is the most immediately binding law. It applies whenever training data, retrieval corpora, prompts, outputs or logs contain digital personal data of individuals in India, and to processing outside India connected with offering goods or services to people in India.
What matters most for AI from 13 May 2027
- Lawful ground: consent under section 6 or a section 7 legitimate use for each purpose, including training.
- Notice (Rule 3): itemised data and specified purposes, which makes vague "to improve our AI" clauses risky.
- Security (Rule 6): encryption, access control, logging and backups across model gateways, vector stores and evaluation datasets.
- Breach (Rule 7): any personal data breach, including prompt-log leaks, is reportable without a threshold.
- Retention (Rule 8(3)): one-year minimum retention of personal data, traffic data and processing logs, which shapes LLM observability design.
- Children (section 9): no tracking, behavioural monitoring or targeted advertising directed at under-18s.
- SDFs (Rule 13(3)): once notified, due diligence that algorithmic software does not pose a risk to Data Principals' rights.
- Publicly available data: the Act does not apply to personal data made publicly available by the Data Principal herself or under a legal obligation.
Deep dive: the DPDP Act for AI systems.
IT Rules and synthetically generated information
On 22 October 2025, MeitY published draft amendments to the IT (Intermediary Guidelines and Digital Media Ethics Code) Rules, 2021 that defined "synthetically generated information" and proposed visible labels and metadata for AI-generated content.
What the draft proposed
- Labels covering at least 10% of the visual area, or the first 10% of audio duration.
- Significant social media intermediaries to obtain user declarations on whether content is synthetic and to verify them with technical measures.
What happened next
Final amendments were notified in February 2026. The final text changed several specifics from the draft, so do not rely on the draft's percentages or mechanics; read the notified amendment for the current labelling, declaration and timing requirements.
Duties that already applied
- Rule 3(1)(b) due diligence obligations on intermediaries regarding the content they host.
- Takedown on lawful orders.
- Loss of safe harbour under section 79 of the IT Act for intermediaries that fail to observe due diligence.
For generative AI products, this layer determines labelling, provenance and moderation design. Details: India deepfake and synthetic content rules.
CERT-In directions for AI infrastructure
CERT-In's Directions of 28 April 2022, issued under section 70B(6) of the IT Act, apply to service providers, intermediaries, data centres, body corporates and government organisations. They do not mention AI, but they apply to AI systems as ICT infrastructure.
Core requirements
- Report specified cyber incidents within 6 hours of noticing them.
- Maintain ICT system logs securely for a rolling 180 days within Indian jurisdiction.
- Synchronise clocks to NIC or NPL NTP servers, or sources traceable to them.
- Designate a point of contact for CERT-In.
- VPN, cloud and data centre providers must keep specified subscriber records for 5 years.
Applying them to AI systems
- Treat LLM gateways, model APIs, vector databases and agent tool-call executors as ICT systems whose logs fall within the 180-day requirement.
- Include prompt-injection compromises, model API key leaks and agent misuse in incident classification.
- Reconcile the 180-day CERT-In log rule with DPDP's one-year minimum under Rules 6(1)(e) and 8(3): in practice, the longer period governs logs containing personal data.
Details: CERT-In rules for AI systems.
RBI's FREE-AI framework for banks and NBFCs
In August 2025, a committee constituted by the Reserve Bank of India and chaired by Prof. Pushpak Bhattacharyya released the report on the Framework for Responsible and Ethical Enablement of AI (FREE-AI). It sets out 7 sutras and 26 recommendations across 6 pillars: infrastructure, policy, capacity, governance, protection and assurance.
Key expectations for regulated entities
- A board-approved AI policy.
- An AI inventory or repository of models and use cases.
- Risk-based classification of AI use cases and governance oversight.
- Disclosure to customers when they interact with AI.
- Consumer grievance routes with human escalation.
- AI-specific incident reporting, audits and model validation.
- Cybersecurity for AI systems.
- A tolerant supervisory approach to first-time errors where safeguards exist.
It also proposes shared infrastructure and an AI innovation sandbox.
Binding status
The report's recommendations become binding only when RBI issues directions or circulars implementing them. Check the latest RBI circulars before treating any item as mandatory. Separately, RBI's 6 April 2018 circular on storage of payment system data already requires payment data to be stored only in India.
Details: RBI FREE-AI framework and data localisation for AI workloads.
SEBI's rules for AI in securities markets
SEBI was early to regulate AI use in markets.
- 2019 circulars (January and May) require market intermediaries, market infrastructure institutions and mutual funds to report their AI/ML systems to SEBI quarterly.
- Early 2025 amendment: SEBI amended its Intermediaries Regulations, inserting regulation 16C, so that a regulated person using AI tools is solely responsible for data privacy and security, for the output, and for compliance with applicable laws, regardless of the vendor supplying the tool.
- June 2025 consultation paper on guidelines for responsible usage of AI/ML in the securities market, covering governance, model testing, disclosure to clients, fairness and data security. Check whether final guidelines have been issued.
- Investment advisers and research analysts using AI must disclose its use to clients.
What it means in practice
Regulation 16C removes the "our vendor's model did it" defence. Firms need vendor due diligence, output review, logging and testing proportionate to the use case. Details: SEBI AI/ML rules.
Putting it together: a compliance map for AI builders
| If your AI system... | Check |
|---|---|
| Processes personal data of people in India | DPDP Act and Rules (from 13 May 2027) |
| Generates images, audio, video or text distributed on platforms | IT Rules synthetic content amendments (Feb 2026 text) |
| Runs on infrastructure you operate | CERT-In 6-hour reporting, 180-day logs, clock sync |
| Is used by a bank, NBFC or payment system operator | RBI directions, FREE-AI implementation status, payment data localisation |
| Is used by a broker, fund, adviser or exchange | SEBI quarterly AI/ML reporting, regulation 16C, disclosure duties |
| Is sold to European customers | EU AI Act and GDPR; see EU AI Act for Indian companies |
Five steps for 2026–27
- Build an AI inventory listing models, use cases, data sources and vendors.
- Map each system to the rows above.
- Prioritise DPDP readiness for 13 May 2027.
- Implement labelling and provenance for generative outputs per the notified IT Rules.
- Adopt the MeitY sutras as internal design principles, since future rules are likely to build on them.
Further reading on suhasbhairav.com
Before 13 May 2027
Compliance checklist
Build an AI system inventory
List models, use cases, data sources, vendors and owners for every AI system.
RBI FREE-AI (inventory recommendation)
Map personal data in AI pipelines
Identify personal data in training, retrieval, prompts, outputs and logs.
DPDP Rules, Rule 3(b)(i)
Confirm lawful grounds for AI purposes
Record consent or section 7 grounds for training and inference uses.
DPDP Act, sections 6–7
Label synthetic outputs per the notified IT Rules
Implement labels and metadata as required by the February 2026 amendments.
IT Rules, 2021 (as amended)
Apply CERT-In logging to AI infrastructure
Keep AI gateway and API logs for 180 days in India and sync clocks.
CERT-In Directions, 28 April 2022
Retain personal data logs for one year
Apply the DPDP one-year minimum to logs containing personal data.
DPDP Rules, Rule 8(3)
Check sector directions
Review the latest RBI and SEBI circulars for binding AI requirements.
RBI / SEBI circulars
Adopt the MeitY sutras internally
Use the seven sutras as design principles in AI governance policies.
MeitY AI Governance Guidelines
Answers
Frequently asked questions
Does India have an AI law?
Does India have an AI law?
No. As of 2026 India has no standalone AI statute. MeitY's India AI Governance Guidelines of November 2025 recommend relying on existing laws, including the IT Act, the DPDP Act, consumer protection and sectoral rules, and amending them where gaps appear. Sector regulators such as RBI and SEBI add their own requirements.
Are MeitY's AI Governance Guidelines binding?
Are MeitY's AI Governance Guidelines binding?
No. The guidelines are principle-based policy, setting out seven sutras, six pillars of recommendations and proposed institutions. They do not themselves create penalties. Binding obligations come from laws and rules such as the DPDP Act, the IT Rules and sector regulators' directions, which the guidelines expect to evolve.
Does the DPDP Act apply to AI models?
Does the DPDP Act apply to AI models?
It applies to processing of digital personal data, which includes training data, retrieval corpora, prompts, outputs and logs that relate to identifiable individuals in India. Core obligations on notice, security, breach, retention and children apply from 13 May 2027. Personal data made publicly available by the individual herself is excluded.
Do AI-generated images need labels in India?
Do AI-generated images need labels in India?
Synthetic content labelling is addressed by amendments to the IT Rules, 2021. A draft of 22 October 2025 proposed visible labels and metadata, and final amendments were notified in February 2026 with changes from the draft. Check the notified text for the exact labelling and declaration requirements that apply to your service.
Is RBI's FREE-AI framework mandatory for banks?
Is RBI's FREE-AI framework mandatory for banks?
Not by itself. The FREE-AI committee report of August 2025 contains 26 recommendations, such as a board-approved AI policy and an AI inventory, but they become binding only when RBI issues directions or circulars. Many regulated entities are adopting them early as good practice. Check the latest RBI circulars.
Who is responsible if a vendor's AI tool makes a mistake in securities markets?
Who is responsible if a vendor's AI tool makes a mistake in securities markets?
The regulated person. SEBI's amendment inserting regulation 16C into the Intermediaries Regulations makes a regulated person using AI tools solely responsible for data privacy and security, output and compliance, regardless of the vendor. Due diligence, testing and output review are therefore essential.
Do CERT-In rules apply to AI systems?
Do CERT-In rules apply to AI systems?
Yes, as ICT systems. CERT-In's directions of 28 April 2022 require specified cyber incidents to be reported within 6 hours, logs to be kept for a rolling 180 days within India, and clocks to be synchronised. LLM gateways, model APIs and agent infrastructure are covered like any other ICT system.
Legal references and sources
- India AI Governance Guidelines (MeitY, IndiaAI Mission) — November 2025 (official text)
- Digital Personal Data Protection Act, 2023 — Sections 4–10 (official text)
- CERT-In Directions under section 70B(6), IT Act — 28 April 2022 (official text)
- RBI FREE-AI Committee Report — August 2025 (official text)
- SEBI AI/ML circulars and Intermediaries Regulations — Circulars of 2019; regulation 16C (official text)
This guide explains the law for engineering and business teams. It is not legal advice; confirm how it applies to your organisation with qualified counsel, and check the Gazette for later notifications or amendments.
About the author

Suhas Bhairav
Systems architect and applied AI practitioner who designs production AI systems for Indian and global companies, with a focus on privacy-by-design under the DPDP Act and the EU AI Act.
Related guides
- India AI & sector rulesIndia AI Governance Guidelines (MeitY) Explained for BuildersMeitY's India AI Governance Guidelines (November 2025) explained for builders: the seven sutras, six pillars, proposed institutions, and how to turn them into engineering practice.
- AI & personal dataDPDP Act for AI Systems: Implementation Guide for Indian CompaniesHow the DPDP Act and Rules apply to LLM apps, RAG, agents and ML pipelines in India, with the architecture controls to build before 13 May 2027.
- India AI & sector rulesIndia Deepfake & Synthetic Content Rules (IT Rules Amendments) ExplainedHow India's IT Rules amendments on synthetically generated information work: the October 2025 draft, the February 2026 final notification, platform duties and DPDP overlap.
- India AI & sector rulesRBI FREE-AI Framework: What Banks & Fintechs Must DoWhat the RBI FREE-AI committee report asks of banks, NBFCs and fintechs, what is and is not binding yet, and how it maps onto DPDP obligations from May 2027.
- India AI & sector rulesSEBI Rules on AI/ML in Stock Market Advisors & BrokersWhat SEBI requires from brokers, advisers, mutual funds and market infrastructure using AI/ML: 2019 quarterly reporting, the 2025 sole-responsibility amendment and the June 2025 consultation.
- India AI & sector rulesCERT-In Rules for AI Systems: Incident Reporting & LogsHow CERT-In's April 2022 directions on 6-hour incident reporting, 180-day logs in India and clock sync apply to LLM gateways, model APIs and AI agents, and how they meet DPDP.