Private video or deepfake of you going viral? How to stop copies and get it removed fast

AI-generated editorial illustration by CyberShitty. Not a photograph or a document from the organisations named.

Platforms must now act within two hours when you complain about nude, sexual or morphed images of you, including deepfakes. How to use that rule, how StopNCII and Take It Down block copies with a hash, and how to keep proof.

01 / The first hour

• Do not pay or bargain with anyone threatening to post it. Paying does not buy deletion. Our sextortion guide explains why.
• Save proof before anything is deleted: the link to each post, the account names, screenshots of the post and of any threats, and the date and time. Our evidence fingerprint tool records a SHA-256 hash of each file, in your browser, without uploading it.
• Complain to the platform and say it shows you nude, in a sexual act, or is a morphed or fake image of you. The two-hour rule below applies.
• Block further copies with StopNCII (if you were 18 or older in it) or Take It Down (if you were under 18).
• Report the crime on cybercrime.gov.in, or call 1930 for help filing.
• Tell one person you trust. You have done nothing wrong; the person who shared it has.

02 / Make the platform remove it: the two-hour rule

Confirmed. Rule 3(2)(b) of the IT Rules, 2021, as amended from 20 February 2026 (G.S.R. 120(E)), says a platform must act within two hours of a complaint about content that shows a person's private area, shows them in full or partial nudity or in a sexual act, or is 'in the nature of impersonation in an electronic form, including artificially morphed images'. Before the amendment the deadline was 24 hours.

• Deepfakes are covered. The amended rules define 'synthetically generated information' (audio, images or video made or altered by computer to look real) and say it counts as information under these rules, so an AI-made nude of you falls under the same rule.
• Someone can complain for you. The rule says 'an individual or any person on his behalf': a family member, friend, lawyer or support organisation.
• Every platform must have a way to receive these complaints (Rule 3(2)(c)). Use the report option, and also write to the platform's India grievance officer. The GAC portal lists the grievance contacts of large platforms, including Meta, WhatsApp, YouTube, X, Snapchat, ShareChat and Telegram.
• In your complaint, name Rule 3(2)(b), paste every link, and say what it shows. Keep the acknowledgement.
• If the platform does not act, or refuses: appeal to the Grievance Appellate Committee at gac.gov.in within 30 days. The GAC tries to decide within 30 days. It does not handle money fraud; that goes to 1930.

The separate 'three-hour' rule in the 2026 amendment is for court orders and government notices to platforms, not for a complaint you make yourself.

03 / Block copies with a hash: StopNCII and Take It Down

A hash is a digital fingerprint of a file. These services make it on your own device: the photo or video is not uploaded or seen by anyone. Only the hash goes to the platforms that take part, which then block matching uploads.

• StopNCII.org is run by SWGfL, a UK charity, and was built with Meta. It is free, works from India and has Hindi pages. Use it if the image is of you, you have a copy, it is nude or sexual, and you were 18 or older in it. AI-made nudes of you count. You cannot submit for someone else. Keep the case number and PIN: they cannot be recovered.
• Take It Down (takeitdown.ncmec.org), run by the US National Center for Missing & Exploited Children, is for images or videos taken when you were under 18, even if you are an adult now. It is anonymous, works worldwide and is available in Hindi and other Indian languages.

Know the limits:

• They work only on platforms that have joined. StopNCII's partner page lists Facebook, Instagram, Threads, TikTok, Reddit, Snap, X, Bluesky and Microsoft's Bing, among others. Facebook and Instagram also use Take It Down, and its own participant list includes YouTube, TikTok and Snap. Telegram and WhatsApp are on neither list, and encrypted chats cannot be scanned. Google announced in 2025 that it would start using StopNCII hashes in Search, but it is not yet on StopNCII's partner list.
• StopNCII matches photos with PDQ/PhotoDNA and videos with MD5, according to its FAQ. A cropped, filtered or clipped copy may not match, so hash each version you find, and a few screenshots of a video too.
• They stop new uploads on member platforms. For copies already up, report each link as in sections 02 and 04.

04 / Where to report on each platform

• Facebook and Instagram: use the report option for someone sharing private images; the flow also offers StopNCII and Take It Down. Indian users can write to Meta's India grievance officer through its online form.
• YouTube: report the video for sexual content. If someone used AI to make a video with your face or voice, file a privacy complaint; it must normally come from you or your legal representative.
• X: X bans intimate media shared without consent, including faces digitally put on someone else's nude body. Report the post as an unauthorised intimate photo or video, or use its private information form.
• Telegram: report the message in the app, or email abuse@telegram.org with the t.me links; Indian users can also write to grievance-in@telegram.org.
• Google Search: ask Google to remove results showing real or fake sexual images of you shared without consent. Google tries to remove duplicates as well, but it can only remove the result from Search, not the website itself.
• Any platform: cybercrime.gov.in also has a 'Report Abuse to Social Media' page with links to each platform's reporting route.

05 / Report the crime, and keep proof that holds up

• On cybercrime.gov.in, choose 'Women/Children Related Crime'. Use Report and Track for your own case: it needs your name and an Indian mobile number, and lets you follow the complaint. Call 1930 for help filing it. You can also go to your local police station; see cyber police contacts for every state.
• Evidence for court: under Section 63 of the Bharatiya Sakshya Adhiniyam, 2023, electronic evidence must come with a certificate, and the certificate asks for the hash value of each file and the algorithm used, such as SHA-256. Recording the SHA-256 hash of what you saved, with the link and time, now saves effort later. Our evidence fingerprint tool does this in your browser.
• A StopNCII hash and an evidence hash are different things: the first helps platforms block copies, the second shows that a file you hand to police has not changed.

06 / If the person is a child

• Do not save, screenshot or forward the image or video, not even to show the police or to warn others. Under Section 15 of the POCSO Act, keeping such material without deleting it or reporting it is an offence, and storing it to share is punishable with up to three years in jail.
• Report the link instead: on cybercrime.gov.in under 'Women/Children Related Crime' (the anonymous option is there for child sexual abuse material), to the police, and on the platform.
• If the child is the one in the image, they can use Take It Down from their own device. Take It Down itself asks people not to download or share such images in order to submit them.
• For help and protection, call 1098, the free 24x7 child helpline.

07 / A deepfake of you that is not sexual

A fake video of you promoting an investment, making a statement you never made, or speaking in a cloned voice is not covered by StopNCII or the two-hour rule. But you still have routes:

• Report it on the platform as impersonation. On YouTube, file the privacy complaint for AI-made face or voice (section 04).
• Complain to the grievance officer. Content that impersonates another person is one of the categories platforms must tell users not to post, and since February 2026 a request to remove it must be resolved within 36 hours. If the platform does not act, appeal to the GAC.
• Report it on cybercrime.gov.in, especially if it is being used to cheat people.

For fake celebrity and official videos used to sell 'investments', see our guide to deepfake investment ads, and for cloned-voice emergency calls, our voice clone scam page.

08 / Lower the risk for yourself and your family

• Keep personal photos and videos to people you trust, and review who can see your profile photos and old albums. Deepfake 'nudify' tools work from ordinary clothed photos.
• Do not accept video calls from strangers who move quickly from a friendly chat to a call.
• Under the 2026 rules, platforms that offer AI image tools must use technical measures to stop users making non-consensual intimate images and child sexual abuse material, and must label other AI-made content. If a tool lets someone do this, report the tool to the platform as well.
• Talk to children and parents about it before it happens: if anyone threatens to share an image, tell a trusted adult at once, never pay, and save the threat. Shame is what blackmailers count on.
• Need support? StopNCII lists Indian organisations that help victims, including RATI Foundation (Meri Trustline), Centre for Social Research, Social Media Matters, Red Dot Foundation, Breakthrough India and Responsible Netism. For mental health support, call Tele-MANAS on 14416.

09 / Confirmed, reported and not known

Confirmed (official texts): the two-hour Rule 3(2)(b), the 36-hour and three-hour timelines, the synthetic-content definition and AI-tool duties (IT Rules amendment G.S.R. 120(E), 10 February 2026, in force 20 February 2026, and MeitY's updated rules); the GAC appeal (gac.gov.in); the Section 63 certificate (BSA 2023, Schedule); POCSO Section 15 (2019 amendment); the cybercrime.gov.in reporting options. Confirmed (the services' own pages): how StopNCII, Take It Down, Google, YouTube, Meta, X and Telegram handle reports, and their partner lists as of 5 October 2026, which change over time. StopNCII's hashing methods come from its own FAQ only. Reported: Google's plan to use StopNCII hashes in Search. Not known: how often platforms actually meet the two-hour deadline; no official figures are published. A draft second amendment to the IT Rules (March 2026) was not in force when we checked; it does not change Rule 3(2)(b).