Chaitanya Rohilla v. Union of India and Rajat Sharma v. Union of India (Delhi High Court deepfake PILs)

By late 2023 deepfakes had become a daily problem in India, with fabricated videos of film stars, politicians and journalists circulating on social media. The country had no statute written for them. Victims and the state relied on the Information Technology Act, 2000, the intermediary due-diligence rules framed under it, and general criminal and civil law on impersonation, obscenity and defamation. Whether a platform that hosts a deepfake, or a tool that makes one, can claim the safe harbour that Section 79 gives intermediaries was unsettled. The Supreme Court's reading in Shreya Singhal v. Union of India limits a platform's takedown duty to court or government orders, which is the backdrop to later debates about how fast synthetic media can be taken down.

Two public interest petitions in the Delhi High Court asked the judiciary to push the Union government toward a regulatory response. The first was filed by Chaitanya Rohilla, a Delhi-based advocate, in 2023; The Hindu reported it on 4 December 2023 and it was first listed for 8 January 2024.⁷ The second was filed in 2024 by Rajat Sharma, the journalist (India TV), after a fake video of him went viral.⁴ The two were heard together, and the court's order of 21 November 2024 records both as W.P.(C) 15596/2023 and W.P.(C) 6560/2024.¹

A third petition, by TV Today Network Limited (W.P.(C) 6037/2025), raised the same concern about deepfake apps and was disposed of in May 2025 after the court referred it to the committee process.² A PIL on AI-generated political content during the 2024 Lok Sabha election was also filed in the Delhi High Court; MediaNama reports it was disposed of on 2 May 2024 because the court trusted the Election Commission to act and would not intervene mid-poll.⁸

Chaitanya Rohilla. In a December 2023 interview Rohilla said the IT Rules, 2021 are "not sufficient" to deal with deepfakes and that interim regulation should have followed once the problem was flagged in 2018.⁷ He said AI and deepfakes should be regulated by the legislature or government and that, until then, use should be banned. He suggested a permission-based model, under which a user such as a filmmaker applies to a government authority and states the purpose, and at minimum a visible watermark so viewers can tell real footage from synthetic footage. He also pointed to a specific prayer that platforms be held accountable for deepfakes uploaded to them, and argued that government requests to platforms are "not going to be enough" without such accountability.⁷

Rajat Sharma. The journalist's petition was reported on 8 May 2024. It asks the court to direct MeitY to identify and block access to platforms and mobile applications that enable the creation of deepfakes, and to appoint a dedicated nodal officer within government to deal with them. The court issued notice to the Centre.⁴

TV Today Network. The media company asked the court to direct the Union to frame a comprehensive mechanism under the IT Act or otherwise to check misuse of apps such as deepfake tools, with penal action. By the time it was heard, the bench recorded that the issue was already before the court in the batch led by W.P.(C) 15596/2023.²

The petitions overlap but pull in different directions. Rohilla's is a regulatory-design petition about AI governance, labelling and platform duty. Sharma's is a harm-response petition built around blocking and a point of contact for victims. Both treat the executive, not the court, as the author of the eventual framework.

Early hearings (2024). Rohilla's petition was listed in January and February 2024.⁷¹² A report of a hearing on Wednesday 28 August 2024 recorded the bench as saying there is a "pressing need to control and regulate the spread of deepfakes on internet". It also said that categorising platforms that generate deepfakes as intermediaries under the IT law may not be correct.⁵ That second remark matters for AI governance. If a service creates synthetic content rather than only hosting what users upload, it may not be entitled to the conditional immunity in Section 79.

24 October 2024. The court directed MeitY to place on record what it was doing about deepfakes; the 21 November order refers to a status report filed "in compliance" with those directions.¹

21 November 2024. MeitY's status report described three strands of work. First, research projects on detection: a project on fake speech detection using deep learning (December 2021 to December 2024, outlay Rs 47.846 lakh), and a C-DAC project for detecting deepfake videos and images (January 2022 to March 2024), whose web tool was in testing and whose desktop application "FakeCheck" had been shared with some law-enforcement agencies. Second, an Advisory Group on AI constituted on 25 September 2023, and a sub-committee on AI governance guidelines constituted on 9 November 2023, whose reports were under examination. Third, a Committee on Deepfakes, set up by an office memorandum dated 20 November 2024, the day before the hearing.¹

The bench found merit in the petitioners' contentions and observed that delay in creation, detection and removal of deepfakes causes hardship to the public. It directed that the Union name its nominees to the committee within one week and that the committee (i) examine the petitioners' suggestions, (ii) consider regulations and statutory frameworks in other countries such as the European Union, and (iii) invite and hear stakeholders, including intermediary platforms, telecom service providers, victims of deepfakes and websites that provide or deploy deepfakes, before reporting as soon as possible, preferably within three months. The matter was listed for 24 March 2025.¹

24 March 2025. MeitY filed a further status report. As recorded in a later order, it said the committee had held two meetings and needed further deliberation, and the court allowed three more months.²

7 May 2025. In the TV Today Network petition, a bench of Chief Justice Devendra Kumar Upadhyaya and Justice Tushar Rao Gedela allowed the petitioner ten days to send suggestions to the MeitY committee, disposed of the writ to avoid multiplicity of litigation, and left it free to seek impleadment in the pending batch. The batch was listed on 23 July 2025.²

The committee is an executive body, not a court-appointed commission. Per the court's order of 21 November 2024, its members include representatives of MeitY's Emerging Technologies, Cyber Security and Cyber Law divisions; the Indian Cybercrime Coordination Centre (I4C) of the Ministry of Home Affairs; C-DAC Hyderabad; the Data Security Council of India (DSCI); Amrita Vishwapeetham University; Prof. Balaraman Ravindran of IIT Madras; and a legal representative.¹ The membership listed in the order does not include a judge, a civil-society group or an intermediary. The court instead required the committee to hear platforms, telecom providers, victims and deepfake sites before reporting.

The directions set a process, not an outcome. The court did not tell MeitY to legislate, to mandate watermarks, or to block particular apps. It required that the committee look at comparative regimes, hear affected parties and consider the petitioners' suggestions. That reflects restraint about the courts' role in technology policy, and it also means that how far the petitions advance depends on what the executive produces.

No published report of the committee has been located. The Union's formal rule-making on synthetic media has moved on a separate track. A LiveLaw commentary describes the Information Technology (Intermediary Guidelines and Digital Media Ethics Code) Amendment Rules, 2026 (G.S.R. 120(E)), effective 10 February 2026, as defining deepfakes, shortening takedown timelines from 36 hours to 3 hours for notices from authorised officers or courts, and requiring technical markers on AI-generated content.¹⁰ The record reviewed for this entry does not show whether the committee's work fed into those rules.

Supreme Court. On 16 May 2025 a bench of Justice Surya Kant and Justice N. Kotiswar Singh heard Narendra Kumar Goswami v. Union of India, W.P.(C) 300/2025, a PIL filed after fake videos of Colonel Sofiya Qureshi circulated.⁸ The petition sought rules on labelling and watermarking, 24-hour takedowns, algorithmic audits and a national AI regulator, among other things. The court noted that the Delhi High Court Division Bench headed by its Chief Justice was already hearing connected matters, including W.P.(C) 15596/2023, and that the Union had responded there. It declined to entertain a parallel petition, relegated the petitioner to the Delhi High Court with liberty to seek impleadment, and asked that his suggestions be considered.³ MediaNama reported the bench as calling the issue serious while cautioning against parallel proceedings.⁸

Gujarat High Court. A separate PIL on deepfakes was filed in the Gujarat High Court in 2026 (Vikas Vijay Nair v. State of Gujarat). On 8 September 2026 the Supreme Court issued notice on the Union's transfer petition (T.P.(C) 2597/2026) and stayed further proceedings in the Gujarat case; see the linked entries.⁶¹¹

Personality-rights suits. The Delhi High Court has also granted many orders in suits by individuals against deepfakes of themselves. Rajat Sharma is himself a plaintiff in one of them, Rajat Sharma v. Tamara Doc, a civil suit in which, according to a Bar & Bench report of 12 September 2026, the court permanently restrained defendants from circulating deepfake videos and images of him and misusing India TV's protected material.⁹ That suit is separate from his PIL (W.P.(C) 6560/2024). Reports sometimes conflate the two matters, so the case numbers matter.

The PILs are the main judicial channel through which the question "who regulates deepfakes, and how?" is put to the Union government, as distinct from case-by-case relief for individual victims. Three features stand out for AI governance.

First, the court framed the problem as one of speed: how fast deepfakes are made, detected and removed. That aligns with the later regulatory emphasis on short takedown windows.

Second, it made the executive show its work. The order forced MeitY to disclose its detection research, its Advisory Group on AI and its AI-governance sub-committee, and the dedicated committee was constituted only on the eve of the hearing. The 21 November 2024 order is therefore a court record that sets out dates and outlays for the Union's deepfake-detection and AI-governance work.

Third, it opened the question whether services that generate synthetic media are intermediaries. The August 2024 remark, reported by Bar & Bench, suggests that courts may be reluctant to extend Section 79 safe harbour to generative tools, though no ruling decides the point.⁵

Criticism has focused on pace and accountability. The committee had met only twice in the four months after its formation, against a requested reporting period of three months, and no report has been published.² At the Supreme Court, Justice Kant was reported to have said that regulators move slower than those who make such content, which is a candid acknowledgment of the limits of a committee process.⁸ Petitioners such as Rohilla argue platforms must be held accountable rather than merely asked to cooperate.⁷ A broader criticism, raised in the related litigation over the Sahyog portal and the fact-check unit, is that fast, executive-driven takedown regimes can overreach; the court's direction to hear platforms and victims partly addresses that concern.

As of 15 October 2026, the lead petition (W.P.(C) 15596/2023) and the connected Rajat Sharma petition (W.P.(C) 6560/2024) are treated here as pending. The last verified listing is 23 July 2025, fixed by the order of 7 May 2025.² The outcome of that hearing and any later orders could not be located on Indian Kanoon or in the legal press reviewed, so this entry does not state what happened after July 2025.

Nothing in the sources reviewed shows that the committee has submitted a report to the court, that MeitY has published one, or that the petitions have been disposed of. The Union has since amended the IT Rules in 2026 to address synthetic content, and the Supreme Court has stayed a similar PIL in the Gujarat High Court on a Union transfer petition, so the Delhi proceedings now sit alongside other litigation about regulating deepfakes.¹⁰¹¹ Readers should check the Delhi High Court case status for W.P.(C) 15596/2023 for the current listing.