Deepfakes, whether fake videos, cloned voices or AI-written text, have become enough of a problem that the Government has had to build out a fairly wide response to them. That response mixes legal provisions, regulatory changes, rules for platforms, and public awareness campaigns, all aimed at keeping the digital space safe and accountable.
As part of this, 13 Responsible AI projects focused on deepfake detection have been approved, and more than 11.37 lakh people have gone through 6,650 cyber security awareness workshops under the Information Security Education and Awareness (ISEA) programme.
The legal groundwork
Several existing laws already give the Government tools to act against AI-generated harm or deepfakes. The Information Technology (IT) Act, 2000 penalises identity theft, impersonation, privacy violations and obscene or sexually explicit content, and it lets the Government block unlawful material, order intermediaries to take down illegal content, and authorise police to investigate cyber offences.
The Bharatiya Nyaya Sanhita (BNS), 2023 adds another layer, covering the spread of false or misleading information that causes public mischief, and it can be used to prosecute organised cybercrime involving deepfakes as well.
Then there are the Information Technology (Intermediary Guidelines and Digital Media Ethics Code) Rules, 2021, which require intermediaries to stop users from hosting or sharing unlawful content, deepfakes among them, along with impersonation, misinformation, privacy violations and material harmful to children. Bigger platforms, those with over 50 lakh registered users, have extra obligations: automated detection tools, published compliance reports, appointed grievance officers, and cooperation with law enforcement.
Tighter rules since February
On 10 February 2026, the Government amended the IT Rules, 2021 specifically to deal with synthetically generated information, deepfakes included. The changes call for clear labelling and traceable metadata on permissible AI-generated content, stronger due diligence from platforms, and a requirement that users be made aware of the legal risks of sharing unlawful AI-generated material.
These revised rules name specific categories explicitly: child sexual exploitation material, non-consensual intimate imagery, impersonation and other harmful AI content. Timelines have also gotten shorter. Content removal following a government or court order now has to happen within 3 hours instead of 36. Grievance redressal has sped up too, with a 2-hour window for sensitive categories like nudity and impersonation.
Platforms are also expected to deploy automated tools to stop unlawful AI content from being created or shared in the first place, and if they don’t, they risk losing safe harbour protection under Section 79 of the IT Act, which would open them up to legal liability.
Advisories, victim support and detection tools
Between December 2025 and March 2026, the Government issued several advisories pushing platforms toward stricter moderation, particularly for misleading, defamatory or religiously sensitive AI-generated content. It has also released a Standard Operating Procedure for handling Non-Consensual Intimate Imagery, meant to guide victims, platforms and law enforcement through prompt removal and response. People can raise complaints through Grievance Officers, and unresolved cases can go to the Grievance Appellate Committees.
On the technology side, the IndiaAI Mission’s Safe & Trusted AI pillar has approved 13 Responsible AI projects, including Saakshya from IIT Jodhpur and IIT Madras, AI Vishleshak for audio-visual forgery detection, and IIT Kharagpur’s Real-Time Voice Deepfake Detection System. Separately, platforms like Gyan Bharatam, Gyan Mitram and BharatiyaGPT use LLSM, OCR, HTR and machine translation to preserve India’s manuscript heritage and make it accessible across languages.
Key Takeaway: CERT-In continues to issue advisories and safety material on AI-related threats including Deepfakes, while the ISEA programme’s thousands of workshops have already reached well over a million people, teachers, students, officials and the public alike, with education aimed at helping ordinary users spot and avoid AI-driven fraud and misinformation.
MCQ’s:
1. Which Government initiative has approved 13 Responsible AI projects for deepfake detection?
A. Digital India Mission
B. IndiaAI Mission
C. National Cyber Security Mission
D. National Knowledge Network
2. Under which provision of the Information Technology Act, 2000 can intermediaries lose their safe harbour protection if they fail to comply with due diligence requirements relating to unlawful AI-generated content?
A. Section 43A
B. Section 66C
C. Section 69A
D. Section 79
Read more: Transforming Agriculture: PMKSY Celebrates Over 36 Lakh Hectares of New Irrigation
