India's Anti-Deepfake Law: Fighting 'Weapons of Mass Distortion'
India's anti-deepfake law is moving forward after fake videos targeting a former army chief spread widely enough to require an official government denial. One legal expert called deepfakes "today's weapons of mass distortion" — here's what's already in force, what's newly proposed, and why India's approach looks different from the EU's.
The Trigger: Fake Videos of a Former Army Chief
Manipulated videos appearing to show former Indian army chief Upendra Dwivedi resigning and criticizing military operations circulated widely enough that India's government and its official fact-checking unit had to publicly debunk them, attributing the campaign to accounts linked to Pakistani propaganda networks. Separately, investigators in Ahmedabad uncovered a scheme in which fraudsters allegedly used AI-generated video to get past India's national biometric verification system and obtain loans using stolen identities.
Legal experts describe deepfakes targeting military leadership as functioning closer to psychological warfare than ordinary fraud — the damage happens the moment a fake video goes viral, well before any takedown request can catch up to it.
What's Already Law in India
India doesn't need to wait for new legislation to have deepfake rules on the books — amendments to the IT Rules already took effect earlier this year, and they're notably aggressive by global standards.
What doesn't yet exist is a dedicated criminal statute spelling out specific penalties for creating or distributing a deepfake. That's the gap the newly proposed law is meant to close.
What's Newly Proposed
The proposed law would define clear criminal punishments for deepfakes and related offenses for the first time, rather than relying on existing IT intermediary rules to cover the harm indirectly. It would also formally cover "digital arrest" scams — a rapidly growing fraud pattern where criminals impersonate police or central bank officials over video calls and pressure victims into transferring money under threat of arrest. Under the proposal, convicted offenders could have their assets or bank accounts frozen.
Legal opinion in India is split on the path forward. Some argue a dedicated law will take too long to draft given how fast the threat is moving, and that amending the existing IT Act framework is the more practical near-term fix, since implementation — not the absence of law — has been the real gap. Others argue that intermediary rules were never designed for this problem and that deepfakes need their own legislative category entirely.
India Isn't Moving Alone
The regulatory pattern spreading across Asia looks similar even where the specifics differ.
| Country | Approach |
|---|---|
| India | 3-hour deepfake takedown and mandatory labeling already in force; dedicated criminal statute now proposed |
| Vietnam | Risk-tiered AI system classification plus mandatory deepfake labeling and chatbot disclosure, passed December 2025 |
| Singapore | Online Criminal Harms Act requires platforms to proactively block scams, including deepfakes, from reaching users |
The scale of what's driving this is significant: a United Nations report released this year estimated that online scam losses across East Asia, Southeast Asia, Australia, and New Zealand totaled somewhere between $88.3 billion and $114.1 billion in 2025 alone.
The Core Problem No Takedown Window Fully Solves
Even a 3-hour or 2-hour removal window doesn't undo the damage a viral deepfake causes in its first minutes online. Legal experts tracking India's rules have pointed out directly that harm to public trust and national security happens at the point a video goes viral, not at the point it eventually gets removed. That's true whether the target is a military officer, a company executive, or an ordinary person targeted in a "digital arrest" scam.
This is why detection matters as much as takedown speed. Being able to verify whether a video, voice call, or image is synthetic before it spreads — not just after a court order arrives — is the difference between preventing a crisis and cleaning one up. UncovAI's video detection and real-time detection for calls and meetings are built around exactly that gap: catching manipulated content at the moment of first contact, whether that's a suspicious video call claiming to be from law enforcement or a clip already spreading on social media.
Frequently Asked Questions
Does India already have deepfake laws, or is this all new?
India already has deepfake rules in force through amendments to the IT Rules, including a 3-hour takedown window and mandatory labeling. What's newly proposed is a dedicated law defining specific criminal penalties for deepfakes, which doesn't yet exist as its own statute.
What is a "digital arrest" scam?
It's a scam pattern where criminals pose as police, central bank, or government officials on a video call and pressure the victim into transferring money or facing supposed legal consequences. India's proposed law would formally define penalties for this alongside deepfake offenses.
Why is a 3-hour takedown window not enough on its own?
Content that's designed to go viral often does most of its damage in the first few minutes of circulation, well before any removal request reaches a platform. A takedown addresses the content's continued spread, not the harm already done.
Is India's approach stricter than the EU's?
In terms of takedown speed, yes — India's 3-hour (and 2-hour, for certain impersonation content) windows are considerably faster than comparable removal timelines proposed elsewhere. The EU's AI Act, by contrast, focuses more on disclosure and provenance marking than mandated removal speed.
Detection Before Distribution
Whether you're screening for compliance, protecting against digital-arrest scams, or verifying a video before it spreads, catching a deepfake at first contact matters more than how fast it can be taken down afterward.
Talk to Us →