Someone Made a Deepfake Video of Me: How to Remove It

A customer messages you a link. It is a video of you, your face and your voice, telling people to invest in a scheme you have never heard of. Or it is a clip that makes you say something cruel, or your face pasted onto someone else's body. You did not make it, and people who know you are already sharing it.

The short answer: save evidence of every copy, then report each one through the platform route built for AI or impersonation content, not just a general report. YouTube, TikTok and Meta all have rules that cover realistic AI fakes of real people. If the video is part of a scam, or it threatens or defames you, take it to MCMC or the police in Malaysia as well.

This guide covers deepfakes of a private person or business owner: fake endorsements, scam ads, mocking or defamatory AI videos and face swaps. It is general information, not legal advice.

Last checked: 5 October 2026.

#How Do You Get a Deepfake Video of You Taken Down?

Save the link, the account name and a dated screenshot of each copy. Then report every copy with the matching route: YouTube's privacy complaint for AI content that looks or sounds like you; TikTok's report under its AI rules, which bar using a private person's likeness without consent; and on Facebook and Instagram, Report ad for scam ads and the impersonation report for fake profiles. In Malaysia, add an MCMC complaint for false or offensive content and a police report for scams, threats or defamation. Each platform decides, and labelled parody may stay up.

#Which Kind of Deepfake Are You Dealing With?

The right report depends on what the video does with your likeness. Pick the accurate reason, not the most serious one: reviewers check the content against the reason you chose.

#Save the Evidence Before You Report

Scam ads and fake accounts can vanish and return under new names, and the police and MCMC will want to see what was there. Our guide to collecting evidence before you report harassment goes further if the fakes are part of a wider campaign.

• The full URL of each video, post or ad, copied from the share or copy link option.
• A screenshot of each one that shows the date, the account name and the URL or app screen in the same frame.
• The account, Page or advertiser name and handle, with a screenshot of its profile.
• Timestamps in the video where your face or voice appears, which YouTube asks for.
• A short note of what is false: what you did not say, what you never endorsed.
• Every report reference, email reply and police report number as they arrive.

#YouTube: The Privacy Complaint for AI Content

YouTube treats AI fakes of a real person as a privacy matter. Its privacy guidelines say you can report someone who uses AI or other tools to alter or generate your voice or image without permission, and give face replacement and simulated speech as examples. To qualify, the video should show a realistic altered or synthetic version of your likeness. Start the privacy complaint from that page rather than the general report button.

YouTube says it weighs several things when it reviews an AI likeness complaint:

• Whether the content is altered or synthetic, and whether it is disclosed to viewers as such.
• Whether you can be uniquely identified. YouTube says a first name alone or a fleeting image is unlikely to qualify.
• Whether the content is realistic.
• Whether it contains parody, satire or other public interest value.
• Whether it shows a public figure in sensitive behaviour, such as endorsing a product.

The complaint must come from you, a parent or guardian, or a legal representative, with a few exceptions, and YouTube does not accept privacy complaints from a company. Give the video URL and the timestamps where you appear. YouTube may first give the uploader time to remove or blur the content. For a scam ad shown before a video, its reporting page says to select Sponsored on the ad and choose Report ad.

#TikTok: Unlabelled or Non-Consensual AI Likeness

TikTok's integrity and authenticity guidelines say it requires clear labelling when AI or editing is used to realistically depict people or scenes, and that it does not allow AI generated content that harms individuals. Among the things the guidelines list as not allowed are using the likeness of private figures without consent and AI created likenesses made to bully or harass. Its other rules, including on impersonation, apply to AI content too.

Report the video and choose the reason closest to what it does with your likeness. If a whole account is pretending to be you, report the account for impersonation too. Where the form allows, describe the harm in plain words: a missing label alone is a weaker case than your face used without consent to mislead or mock.

#Facebook and Instagram: Report the Ad, the Page or the Profile

Deepfake endorsements on Meta's apps usually arrive as paid ads. Meta's help page on reporting an ad says you can select Options next to the ad in your feed and choose Report ad. If the ad has scrolled away, search the Meta Ad Library for the advertiser name or a keyword, find the ad and report it from there. Meta says a report may trigger another review of the ad.

If the scam runs from a profile or Page posing as you, report that as impersonation too. Facebook has a help page on reporting a profile or Page pretending to be you, and Instagram's impersonation help page says you can report in the app or, if you have no Instagram account, through a contact form. If the fake uses your business name, see our guide to fake ads using your brand on Meta.

#Google Search: When the Fake Is Sexual

Google has a specific form for fake explicit content. Its page on removing personal sexual content from Search says you can request removal of fabricated content showing you nude or in a sexual act, including deepfakes, when you are identifiable, the content is fake and it was shared without your consent. Google says it does its best to remove duplicates for sexual imagery requests.

Google removes results from its search pages, not the content from the hosting website, so report it on the hosting platform as well. For content its policies do not cover, the same page points to Google's Legal Help Center for requests under local law.

#What Malaysian Law and MCMC Add

Malaysia now places specific duties on licensed platforms, including social media services, about manipulated media. MCMC's FAQ on the Risk Mitigation Code under the Online Safety Act 2025 gives the Code's effective date as 1 June 2026. It says licensed service providers must ensure that generated or manipulated content closely resembling real people or events is clearly labelled, and that paid ads may only be allowed from advertisers or users verified against government-issued records. MCMC's Online Safety Act page lists financial fraud and scams, and harassment or abusive content, among the kinds of content that can be reported under the Act.

These are duties on the platforms, not a promise that any single video comes down. This is general information, not legal advice.

• MCMC: its complaint page lists "Content on menacing, obscene, false, indecent & offensive" under Online Content Problems, with the portal at aduan.mcmc.gov.my. Attach your links, screenshots and platform report references.
• Police (PDRM): the same MCMC page lists PDRM for cheating by personation, threats to life or property and criminal defamation. Bring printed screenshots and links to the station.
• NSRC on 997: if anyone, you or a customer, has transferred money because of the fake, call 997 and the bank as soon as possible, then make a police report.
• A lawyer: for defamation or a court order. MCMC's page lists civil action alongside the police for defamation, and a lawyer can tell you whether you have a case.

#If Your Face Is Selling a Scam: Warn People Directly

A fake endorsement hurts the people who trust you. While reports are pending, pin a short, calm notice on accounts you control: the video is fake, you do not endorse this, and you never ask anyone to transfer money through an ad. Tell staff what to say if customers call. Do not repost the fake itself.

A live video call using a cloned face is a different trick; see our guide to verifying a suspected deepfake video call.

#What Not to Do

• Do not report before you save evidence. If the post goes, your proof goes with it.
• Do not organise friends to mass report or pick a false reason. Reviewers check the content against the rules, not the number of reports.
• Do not send a copyright notice for a video you did not make unless you own the footage it copies. False claims can carry legal consequences.
• Do not message or threaten the creator. It tends to produce more copies.
• Do not try to find out who made the fake by tracking or hacking them, or make a counter deepfake. Leave identification to the police.

#What Platforms May Leave Up

Some reports will come back with no action. YouTube says it weighs public interest, newsworthiness, parody and satire, and keeps the final decision. A clearly labelled joke, a clip where you cannot be identified, or commentary on something you said publicly may stay up. Keep your evidence, and use the Malaysian routes or a lawyer where the content is false, threatening or part of a scam.

#When a Case Review Makes Sense

Everything above you can do yourself, and the official reports are what actually lead to removal. SocialSafe by AwareXone, our recovery and content removal service, can help when the fakes are spread across several platforms at once, when they are tied to impersonation accounts or a hacked account, or when you need your evidence organised into clear, honest reports for each platform, MCMC or your lawyer. We never file false or mass reports, we cannot force a platform to remove anything, and the platform or a court makes the decision. Our content removal page explains what we do and do not handle, and our Trust Center sets out what we never ask for, including your passwords and login codes.

#Frequently Asked Questions