How to Obtain Interim Injunctions for AI Deepfakes and Unauthorised Use of Likeness in India (2026), a Litigant's Practical Guide

Deepfake injunctions in India have moved from theoretical possibility to courtroom reality in 2026, as High Courts in Delhi and Bombay grant urgent ex parte relief ordering the takedown of AI-generated synthetic content that misappropriates an individual’s likeness. For general counsel, celebrities, brand owners and IP litigators, the question is no longer whether Indian courts will intervene but how fast they can be made to act, and what evidence and pleading strategies deliver results. This guide provides the step-by-step litigation playbook: from emergency evidence preservation through to model interim prayers, the interplay with IT Rules 2026 notice-and-takedown obligations, and the enforcement tactics that compel platform compliance within hours.

Executive Summary, Can You Get a Deepfake Injunction in India?

Yes. Indian High Courts have demonstrated a clear willingness to grant interim injunctions, including ex parte orders, restraining the creation, distribution and hosting of AI deepfakes that misuse a person’s likeness without consent. As reported by LiveLaw, the Delhi High Court in May 2026 protected entrepreneur Aman Gupta’s personality rights and ordered the takedown of AI-generated obscene content featuring his likeness. Separately, as reported by the Times of India, the same court ordered the blocking of deepfakes targeting Congress MP Shashi Tharoor’s personality rights.

Industry observers expect these orders to accelerate as generative AI tools become more accessible, making the existing personality-rights jurisprudence a reliable pathway for urgent interim relief for deepfakes. Critically, the speed of relief, often within 24 to 72 hours on an ex parte basis, depends on the quality and completeness of the evidence filed at the threshold stage.

Three tactical takeaways for rights-holders considering an application:

• Preserve evidence immediately. Screenshots with timestamps, URL archives, metadata downloads and forensic hash values form the backbone of any successful ex parte application.
• File on urgency. Draft the petition and supporting affidavit within 12 to 48 hours of discovery; courts are receptive to urgent mentioning where irreparable reputational harm is demonstrated.
• Serve platforms in parallel. Combine the court application with statutory takedown notices under IT Rules 2026 to maximise the speed and breadth of content removal.

When Will Courts Grant Interim Injunctions for Deepfakes and Unauthorised Likeness?

The legal test for interim injunctions in India is well established. Courts assess three cumulative factors: whether the applicant demonstrates a prima facie case, whether the balance of convenience favours the grant of relief, and whether the applicant would suffer irreparable harm that cannot be adequately compensated by damages alone. In deepfake cases, these factors almost invariably tip in the applicant’s favour, the synthetic content is inherently deceptive, the reputational damage is immediate and unquantifiable, and the balance of convenience overwhelmingly supports restraint given the absence of any legitimate interest in distributing fabricated likenesses.

The 2025–2026 High Court practice confirms this trend. Courts have recognised that personality rights, rooted in Article 21 of the Constitution of India, extend to an individual’s name, image, voice and likeness, and that AI-generated synthetic reproductions engage those rights no less than traditional forms of misappropriation. As reported by BananaIP’s analysis, the doctrinal foundation for personality-rights claims against deepfakes draws on both the right to privacy and the commercial right of publicity.

Typical Relief Courts Impose

• Takedown orders. Directing intermediaries (social media platforms, hosting services, search engines) to remove or de-index specified URLs containing the deepfake content.
• Restraining orders. Prohibiting named and unnamed defendants from creating, uploading, sharing or monetising further deepfake content using the applicant’s likeness.
• Disclosure orders. Requiring platforms to disclose subscriber information, IP logs and upload metadata to identify anonymous perpetrators.
• Preservation orders. Mandating that platforms preserve server-side evidence, including upload timestamps and account activity logs, pending the final hearing.

Ex Parte Practice and the Urgency Test

Courts grant ex parte relief where serving the respondent before the hearing would defeat the purpose of the injunction, for example, where the defendant may delete evidence or escalate distribution. In deepfake matters, the Delhi High Court has accepted that the viral nature of synthetic content and the irreversible reputational damage it causes justify dispensing with prior notice. The applicant must, however, make full and frank disclosure in the supporting affidavit and undertake to serve the respondents promptly after the order is granted.

Causes of Action, Which Legal Claims to Plead First?

Selecting the right combination of causes of action is a strategic decision that shapes the interim prayers, the evidence required and the remedies available at the final stage. The choice depends on who the claimant is (celebrity, private individual or brand), what the deepfake depicts and how it is being used.

Right of Publicity and Personality Rights, Pleading Elements

This is the primary claim in most deepfake injunction cases in India. The applicant must establish that they possess a recognisable persona (name, face, voice), that the defendant has used that persona without authorisation and that the use causes or is likely to cause harm, whether commercial, reputational or emotional. Celebrity image rights in India are now firmly protected under this head, as confirmed by the Delhi High Court’s recent orders. Private individuals can also invoke personality rights, though the evidentiary burden on recognisability may differ.

Copyright, When Available

Where the deepfake incorporates or is derived from original photographs, video footage or audio recordings in which the applicant owns copyright, a claim under the Copyright Act, 1957 provides an additional ground. This is particularly effective where the AI model was trained on or directly manipulates copyrighted source material. The advantage of a copyright claim is the established statutory framework for interim relief under Sections 55 and 62 of the Act.

Passing Off and Trademark Claims

If the deepfake falsely implies an endorsement or commercial association, for instance, using a celebrity’s likeness to promote a product or service, passing off and registered trademark infringement claims are available. These are particularly valuable for brand owners whose trade marks are displayed alongside the fabricated likeness.

Tort, Privacy and Defamation

Deepfakes depicting obscene, defamatory or harassing content engage tortious claims for defamation, invasion of privacy and intentional infliction of emotional distress. Where the content is sexually explicit, criminal provisions under the Information Technology Act, 2000 (particularly Section 66E) and the Indian Penal Code may also apply, offering the additional leverage of criminal proceedings running in parallel with the civil injunction application.

Emergency Evidence and Preservation Checklist for Deepfake Injunctions in India

The strength of an ex parte application stands or falls on the evidence filed at the threshold stage. Courts expect the applicant to present a compelling, well-documented record that leaves no reasonable doubt about the existence of the deepfake, the identity of the victim and the harm caused. The following checklist represents the minimum evidentiary package that should accompany every urgent petition.

How to Capture and Preserve Digital Evidence

• Screenshots with timestamps. Capture full-page screenshots of every platform where the deepfake appears, including the URL bar, date/time stamp and any visible engagement metrics (likes, shares, comments). Use browser extensions or notarised screen-recording tools that embed metadata.
• Archived copies. Save the content using web archiving services (such as the Wayback Machine or Archive.today) to create an independent, timestamped copy that cannot be altered by the defendant.
• Original file downloads. Where possible, download the deepfake video, image or audio file itself. Record the cryptographic hash value (MD5 or SHA-256) of the downloaded file to establish chain of custody.
• Platform notification logs. If takedown notices have already been sent to the platform (whether under the IT Rules or the platform’s own reporting mechanism), retain copies of the notices, auto-acknowledgement emails and any responses received.
• Source material comparison. Where the deepfake is derived from existing photographs or footage, exhibit the original alongside the fabricated version to demonstrate the manipulation.

Expert Report Scope and Forensic Tests

While not always required at the ex parte stage, a preliminary forensic expert opinion significantly strengthens the application. The expert should address whether the content exhibits hallmarks of AI generation (artefacts, inconsistencies in lighting or facial geometry, metadata anomalies) and confirm that the depicted individual did not participate in the creation of the content. Courts have accepted affidavits from digital forensics professionals and AI researchers as evidence for deepfake injunctions.

Using Preservation Letters and Preservation Orders

Before filing the petition, instruct counsel to issue preservation letters directly to the relevant platforms and hosting services. These letters, sent via email to the platform’s designated compliance officer under the IT Rules, put the intermediary on notice and create a documented trail of the platform’s knowledge. If the platform fails to act, this correspondence becomes powerful evidence of non-compliance supporting both the injunction application and potential intermediary liability.

Procedures and Drafting, Urgent Pleadings, Sample Prayers and Affidavit Structure

Speed is the defining characteristic of deepfake litigation. The petition and supporting affidavit must be drafted with precision and filed within hours of discovery. The following framework reflects the approach that has secured interim relief in recent High Court matters.

Affidavit Structure for Ex Parte Applications

The supporting affidavit should follow a clear chronological and logical structure:

• Identification of the applicant. Establish the applicant’s identity, public profile, professional standing and the commercial or personal value of their likeness.
• Discovery of the deepfake. Detail when and how the deepfake was first discovered, by whom, and on which platforms. Attach screenshots and URLs as exhibits.
• Description of the content. Provide a factual description of the deepfake, what it depicts, how it misuses the applicant’s likeness, and why it is false or fabricated.
• Evidence of AI generation. Reference forensic indicators or expert opinion confirming the content is synthetically generated.
• Harm suffered or apprehended. Quantify reputational damage, emotional distress, commercial loss or risk of further dissemination. Emphasise the viral nature of digital content and the impossibility of fully reversing distribution.
• Prior takedown attempts. Exhibit any preservation letters, platform notices or IT Rules takedown requests already made, and their outcomes.
• Urgency justification. Explain why ex parte relief is necessary, typically, the risk that the defendant will delete evidence, the content is actively spreading, and serving notice would defeat the purpose of the injunction.

Model Interim Prayers

A well-drafted prayer clause for an interim injunction for unauthorised likeness should seek relief that is both comprehensive and enforceable. The following model prayers reflect current High Court practice:

• Prayer 1 (Restraint): An interim order restraining the defendants, their agents, affiliates and any person acting on their behalf from creating, uploading, publishing, distributing, sharing or monetising any AI-generated or digitally manipulated content depicting the plaintiff’s likeness, voice or persona.
• Prayer 2 (Takedown): A direction to the intermediary defendants to forthwith remove, de-index and block access to the specified URLs and any mirror or cached copies of the impugned content.
• Prayer 3 (Disclosure): A direction to the intermediary defendants to disclose to the plaintiff the identity, registration details, IP addresses and upload logs of the account(s) responsible for uploading the impugned content.
• Prayer 4 (Preservation): A direction to the intermediary defendants to preserve all server-side data relating to the impugned content, including upload timestamps, access logs and account activity, pending the final disposal of the suit.
• Prayer 5 (John Doe): A John Doe / Ashok Kumar order restraining all persons, whether identified or yet to be identified, from reproducing, re-uploading or further disseminating the impugned content or any derivative works thereof.

John Doe and Anonymous Defendant Pleadings

Where the creator of the deepfake is unknown, as is frequently the case, John Doe (Ashok Kumar) orders allow the applicant to obtain relief against unnamed defendants. Indian courts have a well-developed practice of granting such orders in IP matters, and deepfake cases fit squarely within this framework. The petition should identify the unknown defendants by description (for example, “the person or persons who uploaded the impugned content from Account X on Platform Y”) and seek disclosure orders to ascertain their identity.

Suggested Affidavit Exhibits List

• Exhibit A: Screenshots of deepfake content with URLs and timestamps
• Exhibit B: Archived web copies (Wayback Machine / Archive.today)
• Exhibit C: Downloaded deepfake file(s) with hash values
• Exhibit D: Original source photographs or video (for comparison)
• Exhibit E: Forensic expert report or preliminary opinion
• Exhibit F: Preservation letters and platform correspondence
• Exhibit G: Evidence of harm (media coverage, commercial impact, distress)

Notice and Takedown vs Court Injunction, When to Use Both (IT Rules 2026)

The IT Rules 2026 impose specific obligations on social media intermediaries and significant social media intermediaries to act on complaints regarding content that impersonates individuals using AI-generated or manipulated media. As explained by practitioner commentary, these rules include expedited takedown timelines for certain categories of synthetic content. The question for litigants is not whether to use the statutory notice-and-takedown route or seek a court injunction, in most cases, the answer is both, deployed simultaneously.

Tactical recommendation: Issue the IT Rules takedown notice to every platform hosting the content immediately upon discovery, this starts the statutory clock and creates a compliance trail. Simultaneously instruct counsel to draft the ex parte petition. When the matter reaches court, the notice trail demonstrates both urgency and good faith, and any platform non-compliance becomes an additional ground for relief.

Enforcement and Cross-Jurisdictional Issues

Obtaining a court order is only the first step. Enforcement, compelling platforms, ISPs and foreign-hosted services to actually take the content down, is where many litigants encounter friction. A structured enforcement workflow is essential.

Enforcement Timeline and Platform Compliance

Once the interim order is granted, serve it immediately on the intermediaries’ designated compliance officers (as required under the IT Rules). Major platforms typically have Indian compliance teams and respond to court orders within 24 to 72 hours. For platforms that resist or delay, the court’s contempt jurisdiction provides the necessary leverage. Where content is hosted on servers outside India, blocking orders directed to ISPs can restrict access from Indian IP addresses while cross-border enforcement proceeds.

Dealing with Anonymous Hosts and Mirror Sites

Deepfakes frequently resurface on mirror sites, lesser-known platforms and encrypted channels. The John Doe order addresses this by restraining all persons, known and unknown, from re-uploading the content. Counsel should monitor for re-uploads using reverse image search tools and automated content-matching services, and return to court with supplementary applications if the content re-emerges on platforms not originally named in the proceedings.

AI Deepfake Remedies in India, Damages and Strategic Settlement

While interim relief for deepfakes is the immediate priority, litigants should plan for the full spectrum of remedies available at the final hearing:

• Permanent injunction. A final order permanently restraining the defendant from creating or distributing deepfakes using the plaintiff’s likeness.
• Damages. Compensatory damages for reputational harm, emotional distress and commercial loss. Where the deepfake was used for commercial gain (fraudulent endorsements, advertisement revenue), the court may award aggravated or exemplary damages.
• Account of profits. An order requiring the defendant to disgorge any revenue or profit earned from the deepfake content.
• Delivery up and destruction. An order requiring the defendant to deliver up and permanently destroy all copies of the deepfake, including source files, trained AI models and derivative works.
• Costs. Recovery of legal costs incurred in bringing the proceedings.

Settlement leverage: The combination of a court order, platform takedowns and the threat of damages proceedings often motivates early settlement. Rights-holders should consider whether a recorded undertaking (consent order), with the defendant agreeing to refrain from further misuse, destroy all copies and pay a negotiated sum, achieves the practical outcome more efficiently than proceeding to a contested final hearing.

Practical Timeline, From Discovery to Final Hearing

The following timeline reflects realistic timeframes for the ex parte route based on current High Court practice:

Next Steps, Protecting Your Likeness

If you have discovered AI deepfake content that misuses your likeness, or if you represent a client facing this situation, immediate action is critical. Begin by preserving all available evidence, screenshots, URLs, archived copies and metadata, before any takedown removes the material. Issue IT Rules 2026 takedown notices to the hosting platforms in parallel. Then instruct experienced IP counsel to draft and file the ex parte petition and supporting affidavit within hours, not days.

Deepfake injunctions in India are no longer unprecedented, they are a tested and increasingly efficient remedy. The courts have signalled their willingness to act swiftly, but the outcome depends entirely on the quality of evidence, the precision of the pleadings and the speed of the application. To protect your intellectual property across borders or to find specialist counsel, consult the lawyer directory or explore the international intellectual property guide for further resources.

Sources

1. LiveLaw, Delhi HC Aman Gupta (deepfake) coverage (May 2026)
2. Times of India, Delhi HC Shashi Tharoor (deepfake protection)
3. Ahlawat Associates, Deepfake Liability India (Apr 2026)
4. SCC Online, IPR April 2026 Important High Court Judgments
5. BananaIP, Personality Rights Infringement, Deepfakes & Digital Misuse (India)
6. Ministry of Electronics & IT, IT Rules / Intermediary Guidelines 2026
7. Lexology, Judgment Summaries (2025/2026)
8. Khanna & Associates, Practical Guide to Deepfake Takedowns (May 2026)