Deepfake scams target Indians with sextortion, honey trapping, fake CEO orders, and family-emergency fraud. India is projected to see 8 million deepfake images in 2025, up roughly 900% year on year per I4C and industry estimates. A single WhatsApp profile photo is enough raw material to make a convincing fake of your face. If you receive a video or voice clip from a known person that feels off, do not act on it. Call that person back on a known number before transferring money, sharing data, or believing the request. If a deepfake of you is circulating, capture screenshots, file at cybercrime.gov.in, call 1930, and request platform takedown. The Indian government’s February 2026 deepfake takedown rules require platforms to remove flagged content within 3 hours of a court or government order.
Who this is for
Anyone in India with a public photo online, which is effectively everyone with a WhatsApp, LinkedIn, Instagram, or Facebook account. The most targeted profiles in 2026: founders and senior executives whose voices and faces are used to instruct finance teams, college students and young professionals targeted with fake nude content, women in public roles facing reputation attacks, senior citizens targeted with fake video calls from grandchildren, and politicians or public figures targeted with election-period synthetic content. Parents of teenagers and partners of professionals should read this too. The attack does not need your bank account, your password, or your phone access. It needs your face, which is already out there.
Deepfakes vs voice cloning vs morphed photos
These are different attack modalities. We have written separately about AI voice cloning scams in India, which focuses on audio only. This post covers synthetic visual content: video face swaps, AI-generated compromising images, fake video calls, and full body deepfakes. The defence pattern is similar across all of them: assume the content can be faked, verify out of band before acting, and have a reporting plan ready before you need it.
Why this is spiking now
Three things changed between 2024 and 2026. One: the tools became free and easy. Open-source face-swap models that needed a workstation in 2022 now run on a laptop or even a phone. Two: the cost of compute dropped enough that a scammer can generate hundreds of deepfakes a day on rented cloud GPUs. Three: large language models made the social engineering layer (the script, the timing, the tone) much sharper.
The numbers India is now tracking, sourced from I4C briefings, parliamentary statements, and industry projections:
| Metric | Figure |
|---|---|
| Projected deepfake images circulating in India 2025 | ~8 million |
| YoY growth rate in deepfake volume | ~900% |
| 2025 total cybercrime losses to Indians per I4C / MHA | INR 22,495 crore |
| Estimated share of cybercrime cases involving synthetic content | growing fast, not yet officially broken out |
Per the I4C portal at i4c.mha.gov.in and MHA briefings reported in mainstream press, India saw a 24% rise in cybercrime in 2025. Synthetic content is now a base layer technique across multiple scam categories: sextortion, investment fraud, family emergency calls, fake KYC, and CEO impersonation.
The legal infrastructure has begun to catch up. The IT Rules amendment of 15 November 2025 introduced explicit obligations for platforms to label and act on synthetic content. The February 2026 deepfake takedown rules require platforms to remove flagged synthetic content within 3 hours of a court or government order. The Delhi High Court has issued multiple takedown orders against deepfake content involving public figures. India’s draft Digital India Act, when notified, is expected to tighten this further.
Real Indian cases
These are publicly reported in mainstream Indian press.
Rashmika Mandanna deepfake, November 2023. A morphed video swapping Rashmika Mandanna’s face onto another woman in an elevator went viral. The case triggered India’s first major public conversation on deepfakes and the IT Rules amendments that followed. Widely covered across Indian mainstream press at the time.
Sachin Tendulkar deepfake endorsement, January 2024. A deepfake of Sachin Tendulkar appearing to endorse a betting app circulated widely. Similar fake endorsements have targeted Ratan Tata, Virat Kohli, and Aamir Khan. The Tendulkar case was widely reported across Indian press and prompted a public response from Tendulkar himself.
Mumbai CEO fraud, May 2025. A finance executive received a video call from someone appearing to be the CEO instructing an urgent INR 1.5 crore transfer. The CEO was actually in a board meeting. Pattern documented across multiple Indian press reports in 2025.
Bengaluru tech worker sextortion, July 2025. A 28-year-old IT professional received WhatsApp messages with AI-generated nude images of himself, demanding INR 5 lakh. The face was his LinkedIn photo grafted onto explicit content. He filed at cybercrime.gov.in. (The420.in covers the same gang pattern.)
Election period synthetic content, 2024 general elections. ECI and multiple state CIDs flagged deepfake audio and video clips of national political leaders, with multiple FIRs registered. (Indian Express)
Family emergency fake calls, ongoing. Senior citizens have reported WhatsApp video calls showing grandchildren in distress, asking for money for bail or hospital. The clip is short, the connection breaks, the money request follows from a different account.
The common thread: the technology no longer requires expertise. The defence depends almost entirely on the receiver developing a verification reflex.
How a deepfake is made
You do not need to understand the math to spot the attack. The pipeline:
- Collect target images. WhatsApp display picture, LinkedIn headshot, Instagram public posts, Facebook tagged photos, YouTube videos. For voice, a 10 to 30 second clip from a podcast, voice note, or video. Most of what is needed is publicly accessible.
- Feed into a face-swap or voice-clone model. Free apps and open-source models do this in minutes. Some run on a phone.
- Generate the synthetic content. A short video, a still image, a voice message, or a real-time live filter for video calls.
- Wrap in a social engineering script. Urgent, emotional, time pressured. Fake CEO instructing finance. Fake child in trouble. Fake explicit content with extortion.
- Distribute via WhatsApp, Telegram, Instagram DM, email, or live video call. Often paired with a spoofed number, a fresh SIM, or a mule account.
The whole pipeline takes a moderately skilled scammer under an hour per target, less for repeat templates.
How to spot a deepfake
Detection technology is racing the generation technology and the gap is closing. The reliable defences in 2026 are behavioural, not visual. Visual tells worth knowing:
- Mismatched lip sync on plosives (p, b, m) and sibilants (s, sh)
- Unnatural blinking pattern (too frequent, too rare, or oddly timed)
- Blurry hairline, jewellery, glasses, or background where the face meets it
- Skin tone shifts between face and neck, strange shadows that do not match the light source
- Fingers that bend wrong, fuse, or have wrong count
- Voice clones: flat emotional range, strange pauses, background audio that does not match the claimed location
Behavioural cues are more reliable because they survive any technical improvement: urgency that does not allow verification, requests to keep things quiet, money movement to an unfamiliar account, a story that does not fit what you know of the person, and inability to answer a question the real person would know easily.
The verification reflex (the one habit that defeats most deepfakes)
If you take only one thing from this post: build a verification reflex.
When a video, audio, or image arrives showing someone you know asking for money, data, or urgent action, before you do anything, call that person back on the number you already have saved. Not the number that sent the message. Not the number visible in the caller ID. The number you saved earlier. Even better, use a different channel: if the message came on WhatsApp, call on a regular phone line. If the message came on a video call, call on a normal voice call.
For families, agree on a safe word with elderly parents and children. Something simple, not used elsewhere. If a video call shows your grandchild in distress, ask for the safe word. A real grandchild knows it. A deepfake does not.
For finance teams, no transfer goes out on the basis of a single video or voice instruction. The standing rule: every transfer above a threshold needs a callback to a known number, plus a written confirmation, plus the second signatory. This is not bureaucracy. It is the only thing that catches CEO deepfake fraud.
5 red flags that should make you stop
Any one of these is enough to pause and verify.
- Urgency that does not allow verification. “Transfer now, I am in a meeting, do not call.” The whole point of the pressure is to stop you doing the one thing that catches the scam.
- The person sounds like them but the request is out of character. Your CEO does not usually ask for personal vendor transfers. Your son does not usually call about bail. Out of character is the strongest signal.
- Video or audio quality slightly off. Mismatched lip sync, flat voice, blurry edges. Trust the instinct even if you cannot articulate why.
- Contact from a new number, profile, or account. The voice or face is right but the identifier is new. The scammer cannot easily spoof your saved contact, so they use a new one and hope you do not notice.
- Request involves money, OTP, AePS biometric, or KYC data. Any of these in a video or voice message is a scam by default. Verify before you act.
Privacy hygiene before the attack
You cannot remove every photo of yourself from the internet. You can reduce surface area: lock down Instagram, Facebook, and LinkedIn profile visibility; avoid posting children’s faces with full names; decline video calls from unknown numbers; avoid long public voice notes; use two-factor authentication on every account; watermark sensitive presentations. These do not stop a determined attacker. They raise the cost.
What to do if a deepfake of you is circulating
- Capture evidence first. Screenshots of every post, URL, account handle, timestamp, and the synthetic content itself. Save to a local folder.
- File at cybercrime.gov.in. Use the Report Crime related to Women/Child track for sexual content. Use the general track for fraud, defamation, or CEO impersonation.
- Call 1930. The I4C helpline escalates to local cyber cells.
- File a takedown with the platform safety team. Meta, Instagram, X, YouTube. The February 2026 deepfake takedown rules require removal within 3 hours of court or government order. Cite the rules in your request.
- Inform your inner circle proactively. A short message to family, colleagues, and clients warning that synthetic content is circulating about you.
- Consult a lawyer if reputation or career damage is significant. Civil defamation, criminal complaints under BNS and IT Act, and platform-level legal notices may all apply.
- Do not engage the scammer. If extortion is attached, do not pay. The video does not get deleted because you paid; it gets used again.
What to do if you received a suspicious deepfake from someone
If a friend, family member, colleague, or boss appears in a video or voice clip with an urgent request: pause, call back on a known number on a different channel, ask a verification question only you both know, screenshot before responding, and report at cybercrime.gov.in if confirmed scam.
The legal framework that protects you
Deepfake scams trigger multiple Indian laws: BNS 308 (extortion), BNS 336 (forgery, covers AI-generated synthetic content), BNS 351 (criminal intimidation), BNS 356 (defamation), IT Act 66C (identity theft), 66D (cheating by personation), 66E (privacy violation, capturing or transmitting private images), 67 (obscene material), and 67A (sexually explicit material). The IT Rules 2021 as amended in November 2025 and February 2026 place clear obligations on platforms, including the 3-hour takedown window. Use them.
If this happened to you
If you got a deepfake video, voice clip, or image targeting you or a family member, or if a deepfake of you is being circulated, these channels are free and operate 24x7.
- 1930. National cybercrime helpline operated by I4C under MHA. 24x7. Available across India.
- cybercrime.gov.in. Online complaint portal. File within 24 hours of the incident for fastest response.
- Cybersecify WhatsApp helpline: +91 99644 43350. Free verification for citizens. Send the suspicious video, audio, screenshot, or sender details. We tell you whether it is a known scam pattern and what to do next, in plain language.
- Email: contact@cybersecify.com. For longer evidence packages or written queries.
Save the WhatsApp number now. During an active scam, you will not have time to search.
If you are in a mental health crisis after a sextortion or defamation deepfake, AASRA +91 9820466726 (24x7), Vandrevala Foundation 1860-266-2345 (24x7), and iCALL 9152987821 are free, confidential helplines.
You are not the first, and you are not alone
In 2025 Indians lost approximately INR 22,495 crore to cybercrime per I4C. Tens of thousands of those cases involve synthetic content. Senior citizens, college students, finance executives, women in public roles, founders, parents of teenagers, and ordinary professionals are all in the target set. The volume tells you this is industrial, not personal. The scammers run scripts at scale; they did not single you out for some failing of yours.
The shame belongs to the scammer. The response that protects you is to verify before you act, to report fast, to keep helpline numbers saved, and to talk to one trusted person early. Doing those four things turns most deepfake attacks into a story you tell, not a loss you carry.
We also publish related guides: AI voice cloning scams, sextortion first-hour playbook, digital arrest scams, and WhatsApp GhostPairing account takeovers.
Foundational reads. The anchors behind every guide on this site.
- The First Hour After Cyber Fraud in India. What to do in the first 60 minutes after you realise you have been scammed.
- Pause, Verify, Then Act. The universal three-rule defence against every scam type.
- Your Digital Footprint Is the Scam’s Raw Material. Why scammers already know your name, employer, and bank.
Frequently asked questions
What is a deepfake and how is it different from a normal edited photo?
A deepfake is a synthetic image, video, or audio clip generated by AI that swaps a real person’s face, voice, or body onto another scene. Unlike a Photoshopped image, deepfakes can mimic micro expressions, blinking, lip sync, and voice tone with enough realism that a casual viewer cannot tell. In 2026, free apps and open-source models can produce a usable deepfake from a single WhatsApp profile photo in under a minute. India is projected to see 8 million deepfake images circulating in 2025, growing roughly 900% year on year per I4C and industry estimates.
How can I tell if a video or image I received is a deepfake?
Watch for unnatural blinking, mismatched lip sync, blurry hairlines or jewellery, inconsistent skin tone around the jaw and neck, weird shadows, fingers that bend wrong, and audio that does not match throat movement. In voice clones, listen for flat emotional range or unnatural pauses. If a video shows someone you know asking for money or doing something out of character, call that person on a known number before acting. The verification call is the single most reliable defence.
What should I do if a deepfake of me is circulating online?
Capture evidence first. Take screenshots of every post, URL, account handle, and timestamp. File a complaint at cybercrime.gov.in with the Report Crime related to Women/Child track if it is sexual content, or the general track otherwise. Call 1930. Simultaneously file a takedown with the platform safety team (Meta, X, YouTube, Instagram). The Indian government’s IT Rules amendment of November 2025 and the February 2026 deepfake takedown rules require platforms to remove flagged synthetic content within 3 hours of a court or government order.
Is making or sharing a deepfake illegal in India?
Yes, in most malicious uses. Sharing a sexually explicit deepfake violates IT Act Sections 67 and 67A. Identity theft falls under IT Act 66C. Cheating by personation through a computer is IT Act 66D. Criminal intimidation is BNS Section 351 and extortion is BNS Section 308. Forgery, which covers morphed and AI-generated synthetic content, is BNS Section 336. Defamation falls under BNS 356. Multiple sections usually apply to a single deepfake incident.
Can a single photo from my WhatsApp profile be enough to make a deepfake of me?
Yes. Modern face-swap models need only one clear frontal photo to produce a convincing deepfake. A WhatsApp display picture, a LinkedIn headshot, an Instagram selfie, or a wedding photo on Facebook is plenty. This is why the standard defence is not to hide every photo, which is impractical, but to assume any public photo can be misused and to harden the response side: verification habits, family safe words, fast reporting, and platform takedown awareness.