India’s deepfake rules have been in force for four months, yet AI-generated content continues to spread without labels, and platforms show little sign of enforcing them. The Preity Zinta case, in which the Bombay High Court allowed the actor to sue Google and 15 others over AI-generated deepfakes, is only the latest example of an individual turning to the courts rather than relying on the rules.

The Information Technology (Intermediary Guidelines and Digital Media Ethics Code) Amendment Rules, 2026 came into force on February 20, 2026. They require platforms to label synthetically generated information (SGI), make users declare AI-generated content before uploading it, and remove flagged deepfakes within three hours, or within two hours in cases involving non-consensual intimate imagery (NCII).

Here are 10 instances since the rules took effect in which AI-generated content spread without labels, platforms failed to act, or victims had to seek remedies on their own.

• 1. Instagram told an NCII victim the deepfakes broke no rules. A MediaNama excerpt from a Decode and Tattle investigation documented the case of a woman targeted by an anonymous network using AI-generated nude images. The images were embedded into her Instagram reels and tagged to reach her followers. When she first reported them, Meta said the content did not violate its Community Standards. The rules require NCII content to be removed within two hours.
• 2. An AI app carrying CSAM stayed live on app stores. The same investigation found that a1.art, an AI app with more than one crore downloads on Google Play, contained child sexual abuse material (CSAM). Google did not answer questions about the findings, and the app remained available on both Google Play and Apple’s App Store.
• 3. A fabricated BSE CEO video resurfaced after the rules took effect. A deepfake of Bombay Stock Exchange (BSE) Managing Director and Chief Executive Officer Sundararaman Ramamurthy offering stock tips first circulated in January 2026. It resurfaced later, prompting BSE to issue a second advisory on March 8, 2026. BSE said the video had “resurfaced multiple times” despite efforts by law enforcement agencies and platforms to remove it.
• 4. A deepfake of the Finance Minister promoted an investment scam. An AI-generated video of Finance Minister Nirmala Sitharaman endorsing a high-return investment scheme continued to circulate in waves on Facebook and other platforms, prompting the Press Information Bureau (PIB) Fact Check Unit to debunk it again in 2026.
• 5. At least 19 unlabelled AI election videos spread during state polls. A Quint investigation traced at least 19 AI-generated videos circulating between January and April 2026 during the Assam, West Bengal, Kerala, and Tamil Nadu Assembly elections. None of the videos were labelled as synthetic by users or platforms. Official handles of the Bharatiya Janata Party (BJP), Congress, and the Dravida Munnetra Kazhagam (DMK) shared several of them.
• 6. A party AI-resurrected a deceased Chief Minister for campaigning. During the 2026, Tamil Nadu election campaign, Tamilaga Vettri Kazhagam (TVK) released an AI-generated video recreating the voice of late former Chief Minister C.N. Annadurai and showing him endorsing party leader Vijay. The clip went viral in March 2026 before anyone flagged it as synthetic.
• 7. Communal deepfakes targeted a Congress politician in Assam. AI-generated videos portraying Congress politician Gaurav Gogoi in a communal context, several of them showing him wearing a skullcap, spread through social media accounts and online networks without any synthetic-content labels.
• 8. AI clones of a sitting MP fabricated statements praising Pakistan. Around March 2026, a deepfake campaign cloned the face and voice of Member of Parliament (MP) Shashi Tharoor to create videos falsely showing him praising Pakistan. The content spread across platforms without labels before the Delhi High Court ordered X to take it down on May 10, 2026.
• 9. Deepfakes of schoolchildren and teachers circulated on Telegram. In June 2026, allegedly obscene AI-generated images of female students and teachers at a school in Guwahati circulated on social media and Telegram without consent. Some were reportedly shared for money before the Assam Police Cyber Crime Branch opened an investigation. Because the content targeted minors, it fell into the strictest category under the rules.
• 10. A manipulated clip sexualised a veteran actor.On May 29, 2026, a manipulated video that digitally swapped actor Madhuri Dixit’s face onto another performer’s awards-show footage went viral, with many viewers believing it was genuine. X added a Community Note identifying it as AI-generated only after it had already spread, rather than labelling it under its own obligations.

Why it matters: These cases highlight the gap between regulation on paper and outcomes online. Unlabelled AI-generated content continues to circulate, victims still rely on courts or police intervention, and there is little public information about how platforms are implementing the rules. This makes it difficult to assess whether India’s deepfake framework is effectively protecting users, shifting responsibility to platforms, or merely creating new compliance obligations with limited visible impact.

Also read:

• Is Detection A Better Alternative To Labelling Deepfakes? #NAMA
• Govt Looks To Curb Deepfakes With Draft Amendment To IT Rules, 2021
• Parliamentary Panel Seeks Tougher Deepfake Rules, Raises Practical Concerns on Enforcement