As of May 19, any American whose intimate images have been posted online without permission — including fake nude images generated by artificial intelligence — can demand that the website take them down within 48 hours. If the site refuses, the Federal Trade Commission can fine it up to $53,088 for every single violation. There is no cap on how many violations one company can rack up.
The new power comes from the TAKE IT DOWN Act, a federal law President Donald Trump signed in the Rose Garden one year ago. The criminal penalties for people who post the images started immediately. But the part of the law that forces websites to act — the 48-hour clock — only became enforceable last week. To file a complaint, victims can now go directly to a new FTC website: TakeItDown.ftc.gov.
The law passed Congress almost unanimously. The Senate approved it without a single objection. The House voted 409 to 2. First Lady Melania Trump championed the bill publicly. And yet some of the country's leading civil liberties groups, including the Electronic Frontier Foundation and the Center for Democracy & Technology, argue the law as written will pressure websites to delete a great deal of perfectly legal speech along with the harmful images it targets.
"It should not take a sitting senator getting on the phone to take these down." — Senator Ted Cruz, the bill's lead Senate sponsor
What the law actually does
The TAKE IT DOWN Act does two things. The full name is a mouthful — "Tools to Address Known Exploitation by Immobilizing Technological Deepfakes on Websites and Networks Act" — but the structure is simple.
First, it makes it a federal crime to knowingly post sexual images of a real, identifiable person without their consent. That includes both authentic photos and what the law calls "digital forgeries." These are AI-generated images designed to look like a real person doing something they did not do — the legal term for deepfakes. Posting deepfake nudes of an adult can now carry up to two years in federal prison. Posting them of a minor can carry up to three.
Second, the law creates a takedown system for victims. To use it, a person who has been depicted (or a parent acting for a child) sends a written notice to the website. The notice has to include four things: a signature, enough information for the site to find the image, a brief good-faith statement that the image was posted without consent, and the sender's contact information. The site then has 48 hours to remove the image and make a reasonable effort to take down any identical copies it knows about.
The law applies to what it calls "covered platforms" — essentially any website or app that hosts content posted by users. Social media, messaging apps, image hosts, video sites, dating apps, and gaming services are all in. Email providers and home internet companies are explicitly out.
Why a federal law was needed
The law's political momentum came from two cases involving teenage girls.
In 2023, a male classmate of Elliston Berry, then 14, used a free AI "nudify" app — a tool designed to fabricate nude images of real people — to create explicit images of her and post them on Snapchat. By Berry's own account, the accounts that spread the images remained live on Snapchat for nine months. They were eventually removed within an hour, but only after Senator Ted Cruz of Texas, who would go on to sponsor the bill, personally called the company. Around the same time, a similar incident hit Francesca Mani, a 15-year-old in New Jersey. Both girls testified before Congress.
Then, in January 2024, a graphic AI-generated image of Taylor Swift was viewed 47 million times on X over 17 hours before being removed, according to research published by The Conversation. The image violated X's own rules.
The scale of the problem is documented in several recent studies. A 2023 report from the security firm Home Security Heroes identified 95,820 deepfake videos circulating online — a 550% jump from 2019. Of those videos, 98% were pornographic, and 99% of the people depicted were women. A 2025 study from the child-safety nonprofit Thorn, surveying 1,200 young people, found that one in 17 said someone had used technology to create a deepfake nude of them.
Before the federal law, victims had a patchwork of remedies that mostly did not work. Forty-eight states had laws against revenge porn. Only 32 specifically addressed sexual deepfakes as of early 2025, according to the policy tracker Ballotpedia. A victim whose abuser lived in another state often had no usable legal path. Suing under existing federal civil law was, in the words of the Senate Commerce Committee, "time-consuming, expensive, and may force victims to relive trauma."
The free-speech worry
Almost every civil liberties group that opposes the law also says, explicitly, that they support its goal. Their concern is about how the takedown system is built.
The system is modeled on the Digital Millennium Copyright Act — the 1998 law that lets copyright holders demand removal of pirated material. But there are three differences that legal experts have flagged. Under the copyright law, a person filing a takedown request has to sign a sworn statement, under penalty of perjury, that they own the rights. Under the new deepfake law, they only have to make a "good faith" statement. The copyright law has a formal counter-notice process that lets a wrongly accused user get content restored. The new law does not. And the copyright law has no fixed deadline; the new law gives platforms 48 hours.
The result, critics argue, is that a website facing a takedown request has every incentive to take the material down without investigating. Becca Branum, deputy director of the Free Expression Project at the Center for Democracy & Technology, put it bluntly to the news outlet CyberScoop: "If you think there's any given post where if you ask an attorney, 'Is it worth $53,000 for me to keep this post up,' the answer is always going to be take it down."
The Electronic Frontier Foundation has pointed to a specific reason for concern about misuse. In his joint address to Congress on March 4, 2025, President Trump told lawmakers he planned to use the law personally: "I'm going to use that bill for myself too if you don't mind, because nobody gets treated worse than I do online, nobody." The quote, confirmed by NPR's official transcript, is the strongest piece of evidence civil-liberties groups cite that the law could be used to silence critics, journalists, or satirists.
As of this week, no constitutional challenge to the law has been filed in federal court. Legal commentators across the political spectrum expect one.
Does it actually work?
Early evidence is mixed and incomplete. A preliminary research paper from Princeton's Center for Information Technology Policy, published in May, found that the supply and demand for AI-generated nonconsensual images actually grew on the platforms researchers studied during 2025. Criminal penalties alone did not appear to deter creators. The paper has not yet been peer-reviewed, and its authors note that the platform takedown rules only began enforcement on May 19, 2026. The real test is just starting.
The FTC has opened that test cautiously. On enforcement day, Chairman Andrew Ferguson sent warning letters — not enforcement actions — to 15 of the country's largest platforms: Alphabet, Amazon, Apple, Automattic, Bumble, Discord, Match Group, Meta, Microsoft, Pinterest, Reddit, SmugMug, Snapchat, TikTok, and X. The next day, the agency sent a second round of letters to 12 unnamed "nudify" sites — the AI tools designed specifically to fabricate nude images. The first federal criminal conviction under the law, of an Ohio man named James Strahler II, came in April.
What this means for you
If someone has posted intimate images of you — real or AI-generated — online, you now have a legal tool you did not have a week ago.
Here is the practical sequence. Go to the platform first. Every major website is now required to post a clear notice explaining how to submit a removal request, and most have built dedicated forms. The 19th News maintains an updated catalogue of where to find these forms on the largest platforms. Your written request needs four things: your signature (electronic is fine), a URL or other information that helps the site locate the image, a brief statement that the image was posted without your consent, and your contact information. Keep a copy.
If the website does not remove the image within 48 hours, you can file a complaint with the Federal Trade Commission at TakeItDown.ftc.gov. The FTC — not you — brings the case against the company. You cannot personally sue a platform under this law, but you can sue the person who posted the image, under existing federal and state laws.
If you are the parent of a teenager, the most important thing to know is that the law covers minors and that a parent or guardian can file the request on a child's behalf. Schools, which previously had almost no recourse when a student used an AI app to fabricate nude images of a classmate, can now point families to a federal procedure.
If you run a website, newsletter, forum, or any service where users post content, you are almost certainly a "covered platform," and the obligation to build a notice-and-removal system applies now, not later. The FTC's compliance guidance is published at ftc.gov.
If you publish or share commentary, satire, or journalism online, be aware that the law's broad language, combined with the 48-hour clock, may pull some lawful material down along with the harmful images it targets. There is no penalty for filing a bad-faith request and no formal way to get content restored once removed.
The next year will determine whether the law works as advertised. Watch for three things: whether the FTC moves from warning letters to actual enforcement against a major platform; whether someone files the constitutional challenge that nearly every legal observer is expecting; and whether the supply of nonconsensual deepfakes online begins to fall now that platforms — and not just the people who make the images — face real penalties for letting them spread.
