The platform did not remove your deepfake in 48 hours. Here is what to do.

What the law actually requires

The TAKE IT DOWN Act is Public Law 119-12, signed on May 19, 2025. It has two parts. Since that day, knowingly publishing an intimate image of a real person without consent, real or AI-generated, is a federal crime. Since May 19, 2026, every “covered platform” (a site, app or service that mostly hosts what users post, from social networks to forums and image boards) must offer a notice and removal process, and must act on a valid request within 48 hours: remove the image, and make reasonable efforts to find and remove identical copies.

The Federal Trade Commission enforces the platform side. Ignoring a valid request is an unfair or deceptive practice, with civil penalties above $53,000 per violation. The FTC sent warning letters to about a dozen sites the day after the deadline, and reminder letters to the large platforms.

What the law does not do: it does not cover a deepfake that is not sexual (that is impersonation, handled by platform rules and state law), it does not let you sue the platform yourself, and it does not remove anything on its own. You still have to send the request.

Why your first report probably did not count

Most people report a deepfake the way they report spam: a tap on “report”, a reason picked from a list, and silence. That report lands in an automated queue. The 48-hour clock starts only on a valid request, and the law says what one contains:

• a statement that you are the person depicted, or that you are authorized to act for that person,
• a statement that the image was published without your consent,
• enough information for the platform to find the image: the exact link, not a description,
• a way to contact you, and your signature (typed is fine).

Send those four elements through the platform's dedicated form for intimate images, or to the e-mail address it publishes for that purpose, and keep a copy with the date and time. If you already clicked “report”, send it again this way. The clock starts now.

Send a request that starts the clock

1. Save the proof first. A full-page screenshot with the address and the date, and the exact link of every copy. How to save evidence.
2. Find the right door. Instagram, Facebook, TikTok, Snapchat, X, Reddit and Discord each have a form or an e-mail for intimate images shared without consent. Porn sites usually have a removal page they do not advertise, often linked from the footer.
3. Send the four elements. A short version that works:

   I am the person depicted in the image at [exact link]. It is an intimate image (an AI-generated one) and it was published without my consent. Under the TAKE IT DOWN Act, I request its removal, and the removal of identical copies, within 48 hours. You can reach me at [e-mail]. [Your name], [date and time].

4. Note the date and time you sent it. Hour 48 is counted from there.
5. At the same time, go to Google. Google removes explicit fake images of you from its search results on request, whatever the site does. It takes days, so start today.

Hour 49: the platform stayed silent

Report the platform to the FTC at TakeItDown.ftc.gov. The form asks for the platform, the link, the date of your request, and whether the platform even has a removal process. Attach nothing intimate, the dates and the links are enough. You will not get a case worker, and the FTC will not remove the image for you. What you get is a dated record that the platform missed its legal deadline, which is exactly what it does not want.

Then keep going. Send the request again, mention the first one and its date. Report the page to the company that hosts the site (a “whois” lookup gives its name and abuse address). If you have the image, StopNCII.org creates a fingerprint of it on your own device so partner platforms block copies before they are seen. Search your name again on day 3 and day 7. Most content comes down in the second or third round, not the first.

Where the Act stops, and what fills the gap

The person who made it. The Act's criminal side applies to whoever published it. On top, most states have their own law: California Civil Code 1708.86, Texas Penal Code 21.165, New York Penal Law 245.15 and Civil Rights Law 52-c, Florida Statutes 836.13, Pennsylvania 18 Pa.C.S. 3131, Michigan's Protection from Intimate Deep Fakes Act, Rhode Island 11-64-3, South Carolina 16-15-332, among others. Michigan, California and New York let you sue. The DEFIANCE Act, which would add a federal right to sue with at least $150,000 in damages, passed the Senate on January 13, 2026 and is still waiting in the House as of October 27, 2026. The laws, in plain English.

Not sexual. A fake video of you saying something you never said, or a scam ad with your face, is not covered by the 48-hour rule. Report it as impersonation or manipulated media under the platform's rules. Your face in a scam ad: the steps.

Under 18. Take It Down by NCMEC removes and blocks images of anyone who was under 18 in them, free, and the CyberTipline takes the report. Blackmail. Do not pay, stop replying, keep everything, report to the FBI at ic3.gov. The images are almost never sent, because sending them ends the leverage.

General information, not legal advice. Laws differ from country to country: for advice on your situation, talk to a lawyer where you live.

What we do with this rule

Our Emergency Response for the United States is this guide, done for you by a dedicated handler, within the hour, with follow-up for seven days and a dated evidence file at the end. We write the valid request, send it through the right door on every platform, file with Google, report the platform to the FTC at hour 49, and keep going until the content is gone. $299, one time, charged in US dollars. Remove a deepfake of you, starting this hour.