FAQs on Artificial Intelligence Law in India
FAQs on Artificial Intelligence Law in India
What businesses, founders and individuals need to know
Artificial Intelligence has moved from the laboratory to the boardroom. Indian businesses now use AI to screen job applicants, score credit, answer customers, write code, generate marketing content and run entire workflows through autonomous agents. The law has been catching up quickly. Between November 2025 and July 2026, India saw national AI governance guidelines, the notification of the Digital Personal Data Protection Rules, a new regime for deepfakes and AI-generated content, a Government proposal on paying copyright owners for AI training, and the first substantive ruling of the Delhi High Court on AI and copyright in ANI Media Pvt. Ltd. v. OpenAI.
As technology lawyers in Delhi, we are approached by start-ups, platforms, investors and enterprises with similar questions almost every week. We have brought together answers to these frequently asked questions under the Information Technology Act, 2000 (“IT Act”), the IT (Intermediary Guidelines and Digital Media Ethics Code) Rules, 2021 (“IT Rules”), the Digital Personal Data Protection Act, 2023 (“DPDP Act”), the Copyright Act, 1957 and the latest policy documents of the Government of India.
5 Nov 2025 | 20 Feb 2026 | 24 Jul 2026 | May 2027 · NEXT |
MeitY releases the India AI Governance Guidelines; DPDP Rules notified days later | Amended IT Rules on labelling of AI-generated content come into force | Delhi High Court decides interim application in ANI v. OpenAI | Core DPDP obligations on consent, notice and security take effect |
CONTENTS
Part A The legal framework (Q1–2)
Part B Deepfakes and AI-generated content (Q3–5)
Part C Intellectual property (Q6–8)
Part D Data protection (Q9–10)
Part E Liability, contracts and day-to-day use (Q11–14)
Part F Getting started (Q15–17)
PART A The legal framework
- Does India have a dedicated law on Artificial Intelligence?
No. India has not enacted a standalone AI statute comparable to the European Union’s AI Act. The Government has deliberately chosen to govern AI through existing laws, supplemented by targeted amendments and guidance.
The central policy document is the India AI Governance Guidelines, released by the Ministry of Electronics and Information Technology (“MeitY”) under the IndiaAI Mission on 5 November 2025. The Guidelines contain:
- Seven guiding principles (“sutras”) for safe, trusted and responsible AI.
- Six pillars of recommendations: infrastructure, capacity building, policy and regulation, risk mitigation, accountability and institutions.
- An institutional structure led by an AI Governance Group, supported by a Technology & Policy Expert Committee and the AI Safety Institute, with sectoral regulators retaining enforcement powers in their own domains.
The Guidelines are not binding law in themselves. However, they signal how regulators and courts are likely to judge whether a business has acted responsibly, and they are already shaping procurement terms and investor due diligence. We have set out the wider framework in our note on the Legal Framework for Artificial Intelligence (AI) Business in India.
- If there is no AI Act, which laws actually apply to AI businesses?
A fairly wide set. The principal ones are:
Law | What it governs in AI |
IT Act, 2000 and IT Rules, 2021 (as amended in 2026) | Platforms, safe harbour, takedowns and AI-generated content |
DPDP Act, 2023 and DPDP Rules, 2025 | Personal data used to train, fine-tune or run AI models |
Copyright Act, 1957 and Patents Act, 1970 | Training data, AI outputs and AI-related inventions |
Consumer Protection Act, 2019 | AI products, misleading claims and dark patterns |
Bharatiya Nyaya Sanhita, 2023 | Cheating, impersonation, forgery and defamation using AI |
Indian Contract Act, 1872 and Competition Act, 2002 | AI vendor contracts and algorithmic pricing |
RBI, SEBI and IRDAI regulations | Use of AI by regulated financial entities |
Which of these apply depends on what you build, whom you serve and what data you touch. That mapping exercise is usually the first thing we do for an AI client (Technology Law in India: A Comprehensive Guide for Businesses and Startups).
PART B Deepfakes and AI-generated content
- What are the new IT Rules on AI-generated content?
On 10 February 2026, MeitY notified the IT (Intermediary Guidelines and Digital Media Ethics Code) Amendment Rules, 2026, which came into force on 20 February 2026. The key changes are:
- “Synthetically generated information”: a new defined category covering audio, visual or audio-visual content that is artificially or algorithmically created or altered so that it appears real, such as deepfakes, voice clones and face-swaps. Purely text-based AI output, routine editing and certain educational material fall outside it.
- Labelling and provenance: such content must carry a prominent label and embedded metadata or unique identifiers, which platforms must not allow to be removed or suppressed.
- Significant social media intermediaries (more than 50 lakh registered users in India) must obtain user declarations on whether content is synthetic, verify them through technical measures and label confirmed synthetic content before publication.
- Faster action: access to unlawful content must be disabled within three hours of a lawful order, and the grievance resolution period has been cut from fifteen days to seven.
Non-compliance can cost a platform its safe harbour under Section 79 of the IT Act, exposing it to liability for third-party content.
- Someone has made a deepfake of me or my company’s director. What can I do?
There are parallel remedies, and speed matters:
- Report to the platform through its grievance mechanism; the shortened timelines under the 2026 amendments put platforms under real pressure to act quickly.
- Report to the cyber authorities on the National Cyber Crime Reporting Portal (cybercrime.gov.in) or the 1930 helpline, followed by an FIR where appropriate.
- Invoke criminal law: identity theft and cheating by personation using a computer resource (Sections 66C and 66D of the IT Act), privacy violations and obscene content (Sections 66E, 67 and 67A), and cheating by personation, forgery and defamation under the Bharatiya Nyaya Sanhita, 2023.
- Sue for personality rights: the Delhi High Court has repeatedly protected the name, voice, image and likeness of individuals against AI-driven misuse, including through dynamic and “John Doe” injunctions (Anil Kapoor v. Simply Life India & Ors. (2023)). Businesses can also sue for passing off, trademark infringement and defamation.
Where a deepfake is used for fraud, such as a cloned voice of a director instructing a payment, preserve the evidence (URLs, screenshots, metadata, call records) before it disappears. Where an AI-generated inducement was used to obtain money, a complaint of cheating may also lie (Conundrum on Cheating in India: Civil Wrong or Criminal Offence).
- Do we have to tell customers they are dealing with AI?
There is no general statutory duty to disclose every use of AI. However, the position is shifting. Realistic synthetic audio and video must now be labelled on platforms. Consumer protection law prohibits misleading representations, and the dark pattern guidelines target interfaces that deceive or manipulate users. A chatbot that passes itself off as a human agent, or steers users into choices they did not intend, can attract regulatory attention. Transparent disclosure is now the safer default.
PART C Intellectual property
- Is it legal to train an AI model on copyrighted content in India?
This is the most contested question in the field, and the answer is still developing. On 24 July 2026, the Delhi High Court delivered an interim order in ANI Media Pvt. Ltd. v. OpenAI OpCo LLC [CS(COMM) 1028/2024], the first substantive Indian ruling on the point. The Court:
- Jurisdiction: rejected OpenAI’s objection that Indian courts lacked jurisdiction because its training and servers were located abroad.
- Fair dealing: held, at the prima facie stage, that storing ANI’s works to train the models underlying ChatGPT fell within Section 52(1)(a) of the Copyright Act.
- Outputs: found that ANI had not shown that ChatGPT memorised or substantially reproduced its articles in its responses.
- Relief: declined an interim injunction, observing that one would cause irreparable injury to OpenAI and to the public at large.
This is an interim view; the suit continues on merits and an appeal remains possible. Separately, on 8 December 2025 the Department for Promotion of Industry and Internal Trade (“DPIIT”) released a working paper proposing a hybrid statutory licensing model: a mandatory blanket licence permitting AI training on lawfully accessed works, with royalties payable once the AI system is commercialised, at rates fixed by a Government-appointed committee and collected through a central body. That proposal remains at the consultation stage. Developers should therefore continue to document training-data provenance, licences and opt-out compliance.
- Who owns content generated by AI? Can it be copyrighted?
Indian copyright law is built around human authorship. For computer-generated works, Section 2(d)(vi) of the Copyright Act treats the author as “the person who causes the work to be created”. An AI system itself cannot be an author. Whether a user who merely types a prompt has “caused” the work is unsettled; the greater the human creative contribution in selecting, arranging and editing the output, the stronger the claim. DPIIT has indicated that Part II of its working paper will address the status of AI outputs. Until then, AI vendor terms, employment agreements and client engagement letters should clearly allocate ownership of AI-assisted work.
- Can an AI-based invention be patented?
Potentially, but with care. Section 3(k) of the Patents Act, 1970 excludes a “computer programme per se” and algorithms from patentability. AI-related inventions are examined as computer-related inventions, and applicants must show a technical effect or technical contribution beyond the algorithm itself. Indian patent practice contemplates a natural person as the inventor, so an AI system cannot be named as one.
PART D Data protection
- How does the DPDP Act affect AI development and deployment?
Significantly. The DPDP Act applies whenever digital personal data is processed, including to train, fine-tune or operate an AI model:
- Lawful basis: processing generally requires free, specific, informed and unambiguous consent, or one of the limited “legitimate uses”. Data collected for one purpose cannot simply be repurposed to train a model.
- Publicly available data: the Act does not apply to personal data that the individual has made publicly available. This matters for web-scraped training sets, but it is narrower than it first appears.
- Children’s data: verifiable parental consent is required, and tracking, behavioural monitoring and targeted advertising directed at children are restricted.
- Significant Data Fiduciaries must carry out periodic Data Protection Impact Assessments and audits, and verify that algorithmic software they deploy does not pose a risk to the rights of data principals.
- Penalties can extend up to ₹250 crore per instance for failures such as inadequate security safeguards.
The India AI Governance Guidelines expressly state that the use of personal data for AI training must comply with the DPDP Act. We have discussed compliance strategy in DPDP Act 2023: A Corporate Legal Strategy for Data Protection Compliance in India.
- When do the DPDP obligations actually take effect?
The DPDP Rules, 2025 were notified in November 2025 with phased commencement:
Phase | What comes into force |
14 November 2025 | Definitions, constitution of the Data Protection Board and procedural provisions |
November 2026 | Registration and obligations of Consent Managers |
May 2027 | Core duties of Data Fiduciaries: notice, consent, security safeguards, breach reporting and data principal rights |
With roughly seven months left, AI businesses should already be mapping their training and inference data flows, refreshing privacy notices and consent journeys, and building breach-response processes.
PART E Liability, contracts and day-to-day use
- If an AI system makes a mistake that causes loss, who is liable?
Indian law does not treat AI as a legal person, so liability falls on the persons and companies involved. Depending on the facts, claims may arise:
- In contract, against the AI vendor or the deploying business, subject to warranties, indemnities and limitation of liability clauses.
- In tort, for negligence in designing, testing, deploying or supervising the system.
- Under the Consumer Protection Act, 2019, through product liability actions against manufacturers, sellers and service providers.
- Under sectoral rules, where regulators hold the regulated entity accountable for outcomes, whatever tool produced them.
A business deploying third-party AI cannot outsource accountability. Vendor contracts should address accuracy, data use, IP indemnities, audit rights, incident notification and caps on liability.
- Are contracts concluded by AI agents or chatbots binding?
Generally, yes. Section 10A of the IT Act recognises contracts formed through electronic means, and Section 11 attributes an electronic record to the person on whose behalf an automated system sent it. A business that deploys an AI agent to negotiate, accept orders or make commitments will usually be bound by what the agent does within the authority given to it. Clear limits on agent authority, human approval for high-value actions and well-drafted user terms are the main safeguards.
- Can we use AI for hiring, credit scoring or other automated decisions?
Yes, but these are high-risk uses. India does not yet have a dedicated anti-discrimination law for private-sector algorithms, but biased outcomes can trigger regulatory scrutiny, consumer complaints, sectoral fair-practice requirements and, for State entities, challenges under Articles 14 and 15 of the Constitution. Lenders and other financial entities face growing regulatory expectations on model governance, explainability and human oversight (Digital Personal Data Protection Act, 2023 – Implications for FinTech Businesses). Businesses should test models for bias, keep records of how decisions are reached and offer a route for human review.
- What are the risks of employees using generative AI tools at work?
The main risks are leaks of confidential information and trade secrets, breaches of client confidentiality, disclosure of personal data in breach of the DPDP Act, reliance on fabricated output (“hallucinations”) and uncertainty over ownership of AI-assisted work. The Delhi High Court has cautioned that AI chatbot responses cannot be the basis for adjudicating legal or factual issues (Christian Louboutin SAS v. M/s The Shoe Boutique – Shutiq (2023)). Every organisation should have a written AI acceptable-use policy covering approved tools, prohibited inputs, human review and record-keeping.
PART F Getting started
- What should a business building or deploying AI do first?
- Map your AI use cases and classify them by risk: customer-facing, decision-making, content generation or internal productivity.
- Audit your data: where training and input data came from, what rights you hold in it and whether personal data is involved.
- Prepare for DPDP obligations ahead of May 2027, including notices, consent flows and breach response.
- Review vendor and customer contracts for IP ownership, data use, indemnities, liability caps and AI-specific warranties.
- Implement labelling and metadata controls if you host or generate synthetic audio or video.
- Adopt an internal AI policy, train staff and keep a record of who approved each system and how it was tested.
- What should a person do if targeted by an AI deepfake or impersonation?
- Preserve evidence immediately: links, screenshots, metadata, messages and call records.
- Report the content to the platform and on the National Cyber Crime Reporting Portal or the 1930 helpline.
- Where money has been lost, inform your bank at once so that the transaction can be flagged or frozen.
- Consider an urgent civil action for personality rights and a “John Doe” injunction before the Delhi High Court.
- Why you must have a lawyer in AI-related matters?
AI regulation in India is principle-based and spread across several statutes, regulators and courts, and much of the important case law is emerging from the Delhi High Court. Decisions taken at the design stage, on data sourcing, contract terms and content controls, usually decide liability long before any dispute arises. Early advice from experienced technology lawyers in Delhi can help structure AI products and data pipelines lawfully, draft and negotiate AI licensing and SaaS agreements, prepare DPDP compliance programmes, secure IP in AI-assisted work and act swiftly against deepfakes and impersonation. More on our work is on our Technology practice page, and you can reach us through Connect with us.
ADDITIONAL READINGS
- Technology Law in India: A Comprehensive Guide for Businesses and Startups: https://equicorplegal.com/technology-law-in-india-a-comprehensive-guide-for-businesses-and-startups/
- DPDP Act 2023: A Corporate Legal Strategy for Data Protection Compliance in India: https://equicorplegal.com/dpdp-act-2023-a-corporate-legal-strategy-for-data-protection-compliance-in-india/
- Digital Personal Data Protection Act, 2023 – Implications for FinTech Businesses: https://equicorplegal.com/digital-personal-data-protection-act-2023-implications-for-fintech-businesses/
DISCLAIMER
This article is for information purposes only and should not be taken as legal advice.
To know further details, clarification, assistance or any advice on the use, development or deployment of Artificial Intelligence, data protection compliance, or any legal issues arising from deepfakes or AI-generated content, you may connect with us at admin@equicorplegal.com / 08448824659 and visit www.equicorplegal.com
The law is stated as of October 15th, 2026.