Turn on YouTube's likeness detection
YouTube has a tool that scans for videos where your face appears to be altered or generated by AI, so you don't have to search for fakes yourself.
- You need to be 18 or over and the Channel Owner or Manager. It's experimental and not yet available in every country.
- Setup asks for a government ID and a short selfie video, which YouTube uses as the reference for your face. Verification can take up to 5 days.
- Matches appear in YouTube Studio under Content detection, then Likeness. For each one you can request removal, file a copyright claim or archive it. Official help page: YouTube likeness detection.
Get AI fakes removed on YouTube
If someone used AI to make a realistic video that looks or sounds like you, YouTube lets you ask for it to be removed through its privacy complaint process.
- The video must show a realistic altered or synthetic version of you, and you must be identifiable in it.
- The complaint has to come from you or your legal representative. You can message the uploader first, but you don't have to.
- YouTube weighs things like parody and satire, so not every request leads to removal. Official help page: Request removal of AI-generated content.
Report fakes on TikTok and Instagram
Both platforms label AI content and let you report accounts that pretend to be you. Report quickly, before a fake spreads.
- TikTok's rules ask creators to label realistic AI-generated content, and TikTok can remove, restrict or label content that breaks them. See TikTok's Integrity and Authenticity rules.
- On Instagram, you can report an account that was created pretending to be you, from the account's profile or through Instagram's help center. See Instagram's impersonation help.
- Meta labels content it can detect as AI-made, but it says not all AI content carries the signals needed to spot it. Don't rely on labels alone. See how Meta labels AI content.
Act fast on intimate deepfakes
Sexual deepfakes are the most harmful kind, and there are now dedicated tools and laws for them. You don't have to share the images with anyone to get help.
- StopNCII.org creates a digital fingerprint (a hash) of the images on your own device. Only the hash is sent, never the image, and partner platforms use it to find and remove matches. See how StopNCII works.
- In the US, the TAKE IT DOWN Act requires covered platforms to have a removal process and to take down intimate images, including AI-made ones, and known copies within 48 hours. The Federal Trade Commission (FTC) enforces this from May 19, 2026. See FTC guidance on the TAKE IT DOWN Act.
- In the UK, creating a sexually explicit deepfake of someone without consent became a criminal offence on 6 February 2026, under the Data (Use and Access) Act 2025. See section 138 on legislation.gov.uk.
Know your rights where you live
Laws on AI likeness are changing fast. A few places already protect your voice and face directly, and more are coming.
- Tennessee's ELVIS Act, signed in March 2024, added voice to the state's protections for a person's likeness. See the Tennessee governor's announcement.
- The US NO FAKES Act, which would create a federal right over your voice and likeness, was reintroduced in 2026 but isn't law yet.
- Denmark's culture minister introduced a bill on 8 October 2026 to protect people's appearance and voice from realistic digital imitations, aiming for it to take effect around the new year. See the Danish Ministry of Culture announcement (in Danish).
Check suspicious files and label your own AI content
Some AI content now carries hidden markers you can check. Labeling your own AI content also keeps your channel in good standing.
- Google's SynthID adds an invisible watermark to content made with Google AI. You can upload an image, video or audio clip to Gemini and ask whether Google AI created or altered it. See SynthID.
- Content Credentials, based on the open C2PA standard, work like a nutrition label showing how a file was made. See C2PA.
- If you use AI to make realistic content yourself, such as an AI voice or a cloned face, disclose it. YouTube requires disclosure for realistic AI content, including making a real person appear to say or do something they didn't. See YouTube's disclosure rules.
Common mistakes
Report as soon as you find it. Removals are easier before copies spread to other accounts and platforms.
Reposting spreads it further. Post a short text warning instead and link to your real accounts.
Platforms say not all AI content can be detected. Keep checking for fakes yourself.
Realistic AI content must be labeled on YouTube and TikTok. Unlabeled content can be removed.
Your checklist
0 / 6 donePut it into practice
Tools: AI voice and dubbing tools · Can you monetize YouTube with an AI voice?