AI governance frameworks
AI policies, acceptable-use rules, risk assessments, human oversight and board reporting aligned with India's AI Governance Guidelines and global standards.
- AI policy & register
- Risk assessments
- Board oversight
For companies building, buying or using AI: governance frameworks and policies, AI contracts, deepfake and AI-content rules under the IT Rules, data protection, IP in AI training and outputs, and the disputes and liability that follow when AI goes wrong.
Answer a few questions about how you use AI to see which Indian and EU rules apply.
Since 20 February 2026, the amended IT Rules define "synthetically generated information", require platforms that enable AI content to label it prominently, and cut the time to act on government or court takedown orders from 36 hours to 3 hours. India's AI Governance Guidelines (November 2025) rely on existing laws such as the IT Act and DPDP Act rather than a separate AI law, so liability still runs through contracts, consumer, IP and data protection law. For EU-facing businesses, the AI Act's high-risk obligations now start on 2 December 2027.
From data to deployment, each stage carries its own legal exposure. Tap a stage to see where we come in.
Lawful basis under the DPDP Act, copyright and licences for training data, and contracts with data providers.
AI vendor and licence agreements, warranties, IP ownership of outputs, and indemnities for infringement.
AI policy, risk assessment, human oversight, bias testing and board reporting.
Customer disclosures, AI-content labels, consumer law and sector rules such as RBI's.
Deepfake takedowns, data breaches, regulator notices and claims for harm caused by AI.
We act for AI developers, platforms, enterprises using AI, investors, and people harmed by AI.
AI policies, acceptable-use rules, risk assessments, human oversight and board reporting aligned with India's AI Governance Guidelines and global standards.
AI vendor, SaaS and API agreements, model licences, data-sharing terms, output ownership, warranties, liability caps and IP indemnities.
Takedown notices, complaints to platforms and grievance officers, cyber police complaints, and High Court injunctions protecting name, face and voice.
AI-content labelling and provenance, user declarations, 3-hour takedown processes, grievance redressal and safe-harbour protection under Section 79.
DPDP Act compliance for training and deploying AI: consent and notices, children's data, significant data fiduciary duties and cross-border transfers.
Copyright in training data and outputs, licensing content for AI, protecting models and prompts as trade secrets, and infringement claims.
AI in lending, underwriting, fraud detection and customer service for banks, NBFCs and fintechs, aligned with RBI's FREE-AI framework and consumer protection rules.
Claims for harm caused by AI decisions or outputs, product liability and consumer complaints, vendor disputes, and regulator notices.
For platforms and grievance officers: how long do you have to act after an order or complaint under the 2026 IT Rules?
Timelines follow the IT (Intermediary Guidelines and Digital Media Ethics Code) Amendment Rules, 2026, in force from 20 February 2026. Check the order itself for any specific direction, and keep records of when it arrived and what you did.
| Issue | Law or framework | Key point |
|---|---|---|
| Overall approach | India AI Governance Guidelines (Nov 2025) | Seven principles; no separate AI law for now; existing laws apply |
| AI-generated content & deepfakes | IT Act, 2000 and IT Rules as amended in 2026 | Prominent labels on synthetic content; 3-hour takedown on orders |
| Personal data in AI | DPDP Act, 2023 and Rules, 2025 | Consent, notice and security safeguards; most duties apply from May 2027 |
| Harm to consumers | Consumer Protection Act, 2019 | Product liability, unfair trade practices and misleading claims |
| Training data & outputs | Copyright Act, 1957 | Use of copyrighted works for training is being tested in court |
| Selling into the EU | EU AI Act | Content marking from 2 Dec 2026; high-risk systems from 2 Dec 2027 |
Where AI is used, by whom, on what data and with which vendors.
Legal, ethical and regulatory risks ranked for each use case.
AI policy, approvals, human oversight and incident processes set up.
Vendor, customer and data agreements updated for AI risk.
Teams and leadership briefed on rules, limits and red flags.
Takedowns, breaches, complaints and disputes handled fast.
Not yet. MeitY's India AI Governance Guidelines, released on 5 November 2025, set out seven principles and recommend relying on existing laws such as the IT Act and the DPDP Act, with targeted amendments, rather than a separate AI law for now. AI is therefore regulated through IT, data protection, consumer, IP, sectoral and criminal laws.
From 20 February 2026, the IT Rules cover 'synthetically generated information': audio, visual or audio-visual content artificially created or altered to appear real. Intermediaries that enable such content must label it prominently and take reasonable measures against unlawful synthetic content, and all intermediaries must act on government or court takedown orders within 3 hours instead of 36.
Liability depends on the facts and the contracts. The business deploying the AI is usually answerable to its customers under consumer and contract law, and can seek recourse from its vendor under warranties and indemnities. Platforms keep safe-harbour protection for user content only if they follow the IT Rules. Clear contracts and governance records are the strongest safeguard.
Report it to the platform's grievance officer, file a complaint on the National Cyber Crime Reporting Portal, and, if needed, seek an injunction from the High Court protecting your personality rights. Platforms must act quickly on complaints about intimate images or impersonation under the amended IT Rules.
Yes, where personal data is processed. Businesses need a lawful basis, usually consent with a clear notice, must limit use to the stated purpose, protect the data and report breaches. Most obligations under the DPDP Rules apply from May 2027, and significant data fiduciaries have additional duties including checks on algorithmic risk.
The law is unsettled. Indian copyright law has no broad text-and-data-mining exception, and the question of whether training on copyrighted works is permitted is being tested before the Delhi High Court. Licensing training data, or using content you own or that is openly licensed, reduces risk.
Yes, if you place AI systems on the EU market or their outputs are used in the EU. Under the 2026 Digital Omnibus agreement, content-marking obligations for generative AI apply from 2 December 2026 and obligations for stand-alone high-risk AI, such as AI used in hiring or credit, from 2 December 2027.
An inventory of AI uses, approved tools, rules on what data can be entered, risk assessment before new uses, human oversight of important decisions, bias testing, vendor checks, incident response, and clear ownership at management and board level.
Building, buying or using AI, or dealing with a deepfake or AI-related harm? Tell us what you need. We'll explain the rules that apply and the practical next steps.
Disclaimer: As per the rules of the Bar Council of India, Global Vision Law Firm does not solicit work or advertise. This page and its tools give general information based on the Information Technology Act, 2000 and the IT (Intermediary Guidelines and Digital Media Ethics Code) Rules, 2021 as amended in 2026, the India AI Governance Guidelines, 2025, the Digital Personal Data Protection Act, 2023 and Rules, 2025, and the EU AI Act as amended by the 2026 Digital Omnibus agreement, as understood in November 2026. It is not legal advice. The tools are simplified guides; obligations depend on your specific use of AI.
As per the rules of the Bar Council of India, law firms are not permitted to solicit work and advertise. By clicking the "Agree" button and accessing this website, the user fully accepts that you are seeking information of your own accord and volition and that no form of solicitation has taken place by the Firm or its members. Also, the information about us is provided to the user only on his/her specific request and any information obtained or materials downloaded from this website is completely at the user’s volition and any transmission, receipt or use of this site would not create any lawyer-client relationship.
The information provided under this website is solely available at your request for informational purposes only, should not be interpreted as soliciting or advertisement. We are not liable for any consequence of any action taken by the user relying on material / information provided under this website. In cases where the user has any legal issues, he/she in all cases must seek independent legal advice.
Agree Decline