MeitY Deepfake Advisories to Intermediaries (7 November and 26 December 2023)

Deepfakes entered Indian regulatory policy through a single episode. In early November 2023 a manipulated video that placed actress Rashmika Mandanna's face on a sexually suggestive Instagram clip of another woman circulated widely on social media; Mandanna described herself as scared 'for each one of us who today is vulnerable', and Delhi Police registered a First Information Report, reported on 11 November 2023. A man was arrested in January 2024, and said he had created the video to gain Instagram followers.⁸

MeitY had issued deepfake-related advice before: in February 2023, MediaNama reports, it advised LinkedIn, ShareChat and Snapchat on removing deepfake imagery.³ The November episode brought the issue to the top of the agenda. On 6 November Minister of State Rajeev Chandrasekhar posted that platforms have a legal obligation under the IT Rules to prevent misinformation and act on reports, and the Ministry issued its advisory the next day.¹³ The Prime Minister raised deepfakes on 17 November 2023, citing a fake video of himself playing garba and suggesting that platforms add disclaimers where AI is used, and on 18 November Union Minister Ashwini Vaishnaw said that the government would meet platforms and that safe harbour would not protect those who failed to take adequate measures against deepfakes.⁴

On 7 November 2023 the Press Information Bureau announced that the Centre 'today issued advisory to the significant social media intermediaries' to do five things: exercise due diligence and make reasonable efforts to identify misinformation and deepfakes, and in particular information that violates the platform's rules and user agreements; ensure that such cases are acted on expeditiously and within the timeframes in the IT Rules; ensure that users are caused not to host such content; remove any such content when reported within 36 hours of the report; and disable access to the content or information.¹

The release reminded intermediaries that any failure to act under the IT Act and Rules would attract Rule 7 of the IT Rules, 2021 and could make the organisation 'liable to losing the protection' of Section 79(1) of the IT Act.¹ Rule 7 says that where an intermediary fails to observe the Rules, Section 79(1) does not apply and the intermediary is liable under other laws.¹¹

Chandrasekhar called it the 'second advisory within the last six months', said deepfakes were a major violation that harm women in particular and urged victims to file FIRs at the nearest police station.¹ The first MediaNama report said platforms must remove deepfakes within 24 hours; it was later corrected. The 36-hour period applied to removal on a government or court direction, while the 24-hour period then applied to a victim's complaint about content showing nudity or an artificially morphed image under Rule 3(2)(b).³¹¹ (The February 2026 amendment cut those periods to three hours and two hours.)

Between the two advisories MeitY held what it called 'Digital India Dialogues' with platforms.² According to MediaNama's report of 16 December, it met representatives of Reliance Jio, Meta, Google, Telegram, ShareChat, Snap, Samsung and Apple, and asked them to amend their terms of service to refer explicitly to Rule 3(1)(b) and tell users of their obligations. Chandrasekhar told Mint that a further advisory would follow and warned of stricter regulation if companies did not comply.⁵

The second advisory was announced on 26 December 2023. The release said that MeitY had 'issued an advisory to all intermediaries, ensuring compliance with the existing IT rules' and that it was 'the culmination of the discussions' held over a month. The advisory directed that content not permitted under the IT Rules, particularly the categories in Rule 3(1)(b), 'must be clearly communicated to the users in clear and precise language', including through terms of service and user agreements, and expressly informed to the user at first registration and as regular reminders, at every login and when uploading or sharing information. It also said that users must be told of the penal provisions of the Indian Penal Code, the IT Act and other laws that may be attracted by a violation of Rule 3(1)(b), and that terms of service must say that platforms are obliged to report legal violations to law enforcement.²⁶

Chandrasekhar pointed to Rule 3(1)(b)(v), which prohibits misinformation and patently false information, and to the eleven categories of prohibited content in Rule 3(1)(b), saying that the Rules 'comprehensively' addressed deepfakes. He described the advisory as incorporating procedures agreed with platforms in two dialogues, and said that if violations were noted or reported 'the consequences under law will follow'. MeitY, he said, would 'closely observe' compliance in the coming weeks and follow up 'with further amendments' to the Rules or the law 'if and when required'.²

The advisories drew less public criticism than the AI-model advisory of March 2024, because they restated duties that the IT Rules already imposed. Questions nevertheless arose.

Legal basis and scope. Rule 3(1)(b) is a duty to inform users and make reasonable efforts to prevent content from being hosted; it was written for third-party content, not for content an AI system itself generates. MediaNama questioned in November 2023 whether the clause applies to content an AI generates when Chandrasekhar said that bias in Google's Bard breached Rule 3(1)(b).⁷ The advisories did not resolve that question.

Timelines. The first reports of a 24-hour deadline, which circulated widely, mixed up the period for complaints by victims with the period for government and court orders; the advisory itself cited 36 hours.¹³

Enforcement and complaints. The Union encouraged victims to file FIRs, and the Delhi Police case on the Mandanna video led to an arrest within about two months.¹⁸ A public-interest petition by Chaitanya Rohilla (W.P.(C) 15596/2023) was pending in the Delhi High Court. On 21 November 2024 a bench of Chief Justice Manmohan and Justice Tushar Rao Gedela recorded that MeitY had by an office memorandum of 20 November 2024 constituted a committee on deepfakes, and directed that the committee hear intermediaries, telecom providers, victims and deepfake-providing websites and report preferably within three months.⁹

Design. The December 2023 advisory relied on user-facing reminders at login and upload and on changes to terms of service, and MeitY said at the time that binding amendments would follow if compliance proved inadequate.² The October 2025 draft and the February 2026 amendment were the next steps.¹⁰

The 2023 advisories matter as the first instruments through which the Union connected deepfakes to the safe-harbour regime. They established four ideas that recur in later measures. Deepfakes are treated as misinformation or impersonation covered by Rule 3(1)(b). The Union relies on the threat of losing Section 79(1) immunity under Rule 7 rather than direct penalties on creators. Platforms are asked to inform users of penal consequences, so that deterrence reaches the people who generate the content. And the Ministry signalled that advisories would be followed by binding amendments if compliance fell short.¹²

That sequence followed. The March 2024 advisory extended the approach to AI models, MeitY published draft amendments on synthetically generated information in October 2025, and on 10 February 2026 it notified binding rules. MeitY's own explanatory note to the 2025 draft says it had 'earlier issued multiple advisories' to curb deepfakes and that deepfakes had been raised in both Houses of Parliament.¹⁰ The 2026 amendment turned the December 2023 advisory's user-awareness idea into a quarterly notice requirement (Rule 3(1)(c)) and added warnings specific to synthetic-media tools.¹¹

The sources reviewed do not report a formal withdrawal of the advisories, but they have been overtaken in practice by the binding amendment to the IT Rules notified on 10 February 2026 and in force from 20 February 2026, which sets legally enforceable duties on labelling, prevention of unlawful synthetic content and takedown. The amendment's quarterly user notices and Rule 3(1)(ca) warnings go further than the user-communication duties of December 2023. Law-firm commentary treats the advisories of 26 December 2023 and 15 March 2024 as the instruments that preceded the amendment.¹² The sources reviewed do not report any penalty imposed on a platform for non-compliance with the 2023 advisories.