Deepfake Video Check: how to spot a fake AI video, and how to get it taken down in India
Edited and AI-made videos of real people are spreading faster than ever, and one in two Indians says they have been fooled. Run our 60-second checklist on any suspicious video, then learn exactly what India’s 2026 rules force platforms to do, and how fast.
๐ต๏ธ- Spot it: look for faces that blur when they turn, lips out of sync with the words, odd hands and teeth, flat or robotic voices, and clips with no original source.
- The law since 20 February 2026: India’s amended IT Rules require AI-made or altered videos that look real to carry a visible label and permanent metadata. Big platforms must ask users to declare AI content and check those declarations.
- Takedown clocks are shorter: platforms must act on government or court orders within 3 hours (was 36), and on a person’s complaint about intimate, nude or morphed images of them within 2 hours (was 24).
- If you are targeted: report it on the platform, file a complaint at cybercrime.gov.in or call 1930, and save links and screenshots first. Never reshare the fake, even to complain about it.
Is this video fake? 60-second checklist
Watch the video once more, then tick everything you notice. The checker scores the warning signs and tells you what to do next.
Not every fake is made with fancy AI. Many viral fakes are simple edits: a real clip with someone’s face swapped, their clothes altered, or words dubbed over. In September 2026, a woman’s video of bathing in the Ganga at Haridwar was edited to make her appear indecent and shared widely, and she had to go to the cyber cell and ask people to stop forwarding it. Cases like this are why the checklist looks at the source and context, not just the pixels.
The deepfake video rule of thumb: treat any shocking deepfake video as fake until you find the original. A deepfake video spreads because people forward it before checking, and most deepfake video harm happens in the first few hours, which is why India’s new rules now force platforms to act within hours, not days.
How to spot a deepfake video: 10 warning signs
| Look at | What gives a fake away |
|---|---|
| Face edges | Blurring, flicker or a “mask” line around the jaw, hair and ears, especially when the head turns |
| Eyes | Too little or odd blinking, eyes that don’t track, reflections that don’t match the room |
| Mouth | Lips slightly out of sync with the words; teeth that look blurred or merge together |
| Hands | Extra or melted fingers, hands that pass through objects |
| Skin and light | Waxy, too-smooth skin; shadows that fall the wrong way for the light |
| Voice | Flat or robotic tone, no breathing, odd pauses, or an accent that doesn’t match the person |
| Background | Warped lines, text that turns to gibberish, objects that shift between frames |
| Source | No original uploader, posted by a new or anonymous account, or only shared on WhatsApp |
| Message | Shocking, sexual or outrage-bait, and designed to make you forward it fast |
| Coverage | No credible news outlet or the person themselves has posted it |
India’s deepfake rules in 2026, explained
The Ministry of Electronics and IT notified amendments to the IT Rules on 10 February 2026, and they took effect on 20 February. They cover “synthetically generated information”: audio, images or video created or altered by a computer so that it looks real and could be mistaken for a real person or event.
| Rule | What it says |
|---|---|
| Visible label | AI-made or altered content that looks real must carry a prominent, easily noticed label. Audio needs a spoken disclosure at the start. |
| Permanent metadata | Where technically possible, it must also carry permanent metadata and an identifier showing which platform’s tool made it. Platforms must not let anyone strip the label. |
| User declaration | Big platforms (over 5 million Indian users) must ask users to declare if their upload is AI-made, and use tools to check. |
| Banned outright | AI tools must try to block unlawful fakes such as child sexual abuse material, non-consensual intimate images, false documents and deceptive depictions of real people or events. |
| Takedown on orders | 3 hours from a government or court order (was 36 hours). |
| Your complaint about intimate or morphed images | 2 hours (was 24 hours). |
| Grievances | Platforms must resolve complaints in 7 days (was 15). |
| Not covered | Normal edits like colour correction, noise reduction, compression or translation that don’t change what the content shows. |
The early draft (October 2025) wanted labels to cover 10% of the screen; the final rules dropped that number. Critics say the rules still lean heavily on platforms’ own tools and don’t cover many other ways companies use AI on consumers.
How to report a deepfake video in India, step by step
- Don’t reshare it, even to warn people. Every forward spreads the harm.
- Save evidence first: copy the links, take screenshots showing the account name and date, and note where you saw it.
- Report it on the platform using its report or grievance option. If it shows you in an intimate or morphed way, say so: the 2-hour clock applies.
- File a cybercrime complaint at cybercrime.gov.in, or call the helpline 1930. You can also go to your local police station or cyber cell.
- Escalate if ignored. If the platform’s grievance officer doesn’t act, you can appeal to the Grievance Appellate Committee at gac.gov.in.
- If you’re a woman or minor targeted, the National Commission for Women and child-protection services can also help, and police can act under the IT Act’s provisions on privacy and obscene content.
Scammers now use cloned voices and faces too. Read our guide to new WhatsApp and UPI scams to learn the money tricks that often come with them.
Deepfake videos: FAQs
How can I tell if a video is a deepfake?
Look for blurring around the face when the head turns, lips out of sync, odd blinking, strange hands or teeth, and a flat voice. Then check the source: if no credible outlet or the person themselves has posted it, treat it as suspect and reverse-search a frame on Google Lens.
Is making a deepfake illegal in India?
Making or sharing a fake that defames someone, shows them in an intimate way without consent, impersonates them or spreads false information can be an offence under the IT Act and the Bharatiya Nyaya Sanhita. Since February 2026, platforms must also label AI-made content and remove unlawful fakes quickly.
How fast must a platform remove a deepfake of me?
If you complain that content shows you nude, in a sexual act or in a morphed or impersonating image, the platform must act within 2 hours under the 2026 rules. Government and court orders must be followed within 3 hours.
Where do I complain about a deepfake?
Report it on the platform first, then file a complaint at cybercrime.gov.in or call 1930. Keep links and screenshots as evidence, and don’t forward the video.
Do AI videos have to be labelled in India?
Yes. Since 20 February 2026, AI-made or altered content that looks real must carry a clear visible label and, where possible, permanent metadata. Routine edits like colour correction are exempt.
Sources
- MediaNama: MeitY notifies amendments to IT Rules on synthetic media (Feb 2026)
- HLC: India introduces mandatory AI labelling and 3-hour takedowns
- MeitY: explanatory note on the October 2025 draft
- Brut India: Ganga dip video doctored (Sept 2026)
- WITNESS: critique of India’s synthetic media rules
- National Cyber Crime Reporting Portal
This page is general information, not legal advice. The checklist gives a rough risk signal and cannot prove a video is real or fake. If you are a victim, contact the police or a lawyer. WellthRise is not affiliated with any government body or platform.

We write practical guides on money, tax, salary, wellness and India news. Rates, rules and dates come from official sources such as RBI, SEBI, the Income Tax Department, India Code and government notifications, linked in each article with the date we last checked them. We don’t accept payment to change what we write.
Our editorial policyReport an error