TL;DR
To remove a deepfake video, first document the exact URL, a screenshot and the date. Then confirm there is no real underlying original: a real copied video is a DMCA matter, but synthetic content falls under the TAKE IT DOWN Act. Report it to the hosting site with a complete notice (signature, precise URLs, a good-faith statement and contact details); the 48-hour clock starts only once it is complete. Then ask search engines to remove the page and watch for reposts. You don't need to identify the AI tool used, and a real original doesn't need to exist.
If you are reading this because someone made a deepfake of you, you are not at fault and there are clear steps to take. They work for a video, an image or a clip posted anywhere; the order matters. For the wider picture of what we do, see our deepfake removal service.
Why there is a remedy even with no "original" to point to
Finding AI-generated content that depicts you is disorienting in a way an ordinary leak isn't: no real photo or video underneath it was stolen. Someone generated it, or manipulated an existing image into something that never happened. People often assume, wrongly, that this leaves no clean legal remedy, because most of us think of "getting something taken down" as copyright, and copyright needs a real original work to protect.
There is a clean remedy, and it doesn't require a real original. Here's the process, step by step.
Step 1: Document the exact URL
Before doing anything else, capture:
- The exact URL of the page or post: the specific location, not the site's homepage.
- A screenshot of the content and the page around it, including the account or username that posted it, if visible.
- The date and time you found it.
A valid removal notice legally requires precise URLs or information sufficient to locate the content. "It's somewhere on this site" doesn't meet that bar. Content on these sites can also move, be taken down or be reposted without warning, so document it before it changes.
Step 2: Confirm there's no real underlying original
This decides which notice you file. If the content is a real photo or video of you that was copied or leaked without permission, it's a copyright matter and the notice is a DMCA notice. If it's synthetic (a face swap, an AI-generated image or video, or something manipulated to depict an event that never happened), there's no copyrighted original to claim, and DMCA is the wrong tool, or at best a weak, indirect one. This is squarely what the federal TAKE IT DOWN Act was built for.
You don't need to prove how it was made
You don't need to identify the specific AI tool used, or prove the technical process behind it. The statute's threshold is simpler: the content is a non-consensual intimate depiction of an identifiable person that a reasonable viewer would find indistinguishable from authentic — it doesn't require the depicted event to have ever really happened.
Step 3: Report it to the hosting site under the TAKE IT DOWN Act, not DMCA
Public Law 119-12, signed May 19, 2025, requires "covered platforms" to remove qualifying non-consensual intimate imagery, explicitly including AI-generated "digital forgeries," within 48 hours of receiving a complete, valid notice. Covered platforms are services that primarily host user-generated content, or whose regular business involves publishing this kind of content. If you're not sure your situation qualifies, our TAKE IT DOWN Act eligibility checker walks through it in a couple of minutes.
A complete notice needs four things in writing:
- A signature, yours or an authorized representative's on your behalf.
- The precise URL(s) or information sufficient to locate the content.
- A brief good-faith statement that the content is non-consensual.
- Contact information sufficient for the platform to follow up.
Send it to the platform's designated contact for this kind of report, not to the person who posted it. The 48-hour clock starts only once the notice is complete, and an incomplete submission doesn't trigger the deadline, so check that all four elements are present before sending. For the full breakdown of what the statute covers and excludes, see our TAKE IT DOWN Act explainer.
What DMCA can still do here
If the deepfake was built by manipulating a specific photo or video you own the rights to (rather than generated from scratch), you may have a secondary DMCA claim over that underlying source material — but the TAKE IT DOWN Act notice is still the more direct tool for the synthetic content itself, since it doesn't depend on proving ownership of anything.
Step 4: Escalate if the platform doesn't respond
A covered platform's failure to reasonably comply is treated as an unfair or deceptive practice under the FTC Act, and the FTC can enforce against it. If a platform ignores a complete, valid notice past the 48-hour window, that's grounds to escalate. Keep your original notice and any correspondence, because that record is what an escalation would rely on.
Not every site is a "covered platform" under the statute. The exclusions include things like email services and platforms that only publish pre-selected, non-user-generated content. A genuinely obscure or foreign-hosted site may therefore not be bound by the 48-hour requirement even if the content itself qualifies. In that case the fallbacks are a host-level abuse report and, where applicable, a request under a relevant state right-of-publicity or NCII law. These vary by state, so the specifics depend on where you and the poster are located.
Step 5: Request search removal
Even after the content is removed at the source, it can still appear in search results until the page is re-crawled. Google has a dedicated removal request for personal explicit imagery, including AI-generated content, that de-indexes the URL directly instead of waiting on a re-crawl. Our guide to removing leaked content from Google search results explains the mechanics, which apply the same way to synthetic content, and our de-indexing request generator drafts the request itself.
Step 6: Monitor for reposts
Deepfake content that has been removed once can easily resurface, re-hosted on the same platform under a different account or mirrored to a new site, sometimes by whoever generated it in the first place. There's no original file to check against, so catching a repost takes the same identity-based approach as the initial detection: matching your actual face across newly scanned content, rather than searching for a file or filename that may not match the reposted version.
A one-time takedown without follow-up monitoring is a temporary fix. If the same person or account keeps posting, you may not learn a new copy exists until someone tells you.
Quick reference: the whole process
| Step | What to do |
|---|---|
| 1. Document | Screenshot the content and record the exact URL, date, and time |
| 2. Classify | Confirm there's no real underlying original — this points you to the TAKE IT DOWN Act, not DMCA |
| 3. File | Send a complete notice (signature, URL, good-faith statement, contact info) to the platform |
| 4. Escalate if needed | FTC enforcement applies if a covered platform doesn't reasonably comply |
| 5. De-index | Request removal from Google search results directly |
| 6. Monitor | Watch for reposts using identity-based (not filename-based) detection |
When to get help
You can do all of this yourself, and the steps above are written so that you can. Help is worth having when:
- the video has been copied to several sites, so each needs its own notice;
- a site ignores your notice and you need to escalate to its host;
- the same content keeps coming back after removal;
- it is simply too distressing to keep looking at it, which is a good reason to hand the finding and filing to someone else.
If you feel unsafe, or the video comes with threats or demands for money, consider reporting it to the police as well. This is general information, not legal advice.
If you are not sure whether a fake of you exists in the first place, how to find out whether there is a deepfake of you covers searching by face. If you do not have the original photo or video the fake was made from, see removing a deepfake with no original photo.
If you want this handled rather than done manually
Noticeora verifies whether content is genuine or synthetic. On the Autopilot tier ($99/month, which also covers de-indexing and reupload monitoring) it files the correct notice type automatically; the DIY tier ($39/month) gives you detection and a self-filed notice path. Either way, the first useful step is knowing what's actually out there, and a free scan shows you that before you decide how much of this process to take on yourself.
