Resource guide

TAKE IT DOWN Act Rules for Removing AI Deepfakes

How to get AI deepfakes removed under the federal TAKE IT DOWN Act, with covered platforms acting within 48 hours of a valid request.

Last updated October 14, 2026 1647-word guide Editor Ban the Bots

What the TAKE IT DOWN Act Does

The TAKE IT DOWN Act (S.146, Public Law 119-12) makes it a federal crime to knowingly publish nonconsensual intimate images. The law also covers digital forgeries made with artificial intelligence (AI), and it covers threats to publish either an authentic image or a forgery. Since May 19, 2026, covered platforms must remove such images, and known identical copies, within 48 hours of a valid request. The Federal Trade Commission (FTC) enforces that removal duty.

We cover this law because it names AI-made fakes directly, and its removal rules reach them too.

What the Law Makes a Crime

The criminal rules differ by victim age and by image type. Threats to publish an image are a separate crime. The rules below follow Sec. 2 of the enrolled text, which amends 47 U.S.C. 223.

Adult Victims

For an identifiable adult, knowingly publishing an authentic intimate image through an interactive computer service is a crime when all four of these conditions are true.

Publishing a digital forgery of an adult without consent is a separate offense. It uses the same conditions on public exposure, public concern and harm.

Minor Victims

Publishing an intimate depiction of an identifiable minor is a crime when the publisher intends to abuse, humiliate, harass or degrade the minor. It is also a crime when the intent is to arouse or gratify anyone's sexual desire.

How the Law Defines a Digital Forgery

The law defines a digital forgery as an intimate depiction of an identifiable person. The depiction must be created or altered with software, machine learning, AI or other technological means. A reasonable person must find it indistinguishable from an authentic depiction when viewed as a whole. For spotting a digital forgery, see how to spot a deepfake.

Threats and Penalties

Publishing an authentic image or a digital forgery carries a fine and up to 2 years in prison for an adult victim. For a minor victim, the maximum is 3 years. Threatening to publish an authentic image to intimidate, coerce, extort or cause mental distress carries the same penalties. Threatening to publish a digital forgery carries up to 18 months for an adult victim and up to 30 months for a minor victim.

Courts must order forfeiture of the published material and any proceeds. Courts must also order restitution. The law exempts lawful law-enforcement and intelligence activity.

It also exempts good-faith disclosures. These include disclosures to police, in legal proceedings, and for medical, scientific or educational purposes. The law also exempts disclosures made to report unlawful conduct or to seek help. People who publish their own nudity are exempt too. For other rules on deepfakes, see our deepfake laws guide.

What Platforms Must Do Within 48 Hours

The removal duty applies only to covered platforms. The table lists each requirement in the enrolled text.

RequirementWhat the law says
Who must complyA covered platform is a public-facing website, online service or app that mainly provides a forum for user-generated content. A site, service or app that regularly publishes, curates or hosts nonconsensual intimate depictions as part of its business is also a covered platform.
Who is excludedBroadband providers and email. Services built mainly on preselected, non-user-generated content, where chat or comments are incidental.
What a valid request containsA physical or electronic signature. Enough information to identify and locate the image. A brief statement of good-faith belief that the image is not consensual. The requester's contact information.
NoticeThe platform must give clear and conspicuous notice of its removal process, written in plain language.
Removal deadlineAs soon as possible, and not later than 48 hours after a valid request.
Identical copiesThe platform must make reasonable efforts to identify and remove known identical copies.
Start datePlatforms had to set up the process within 1 year of enactment, which made the deadline May 19, 2026.
Good-faith removalA platform is not liable for removing material in good faith, even if the material later proves lawful.
EnforcementA failure to reasonably comply is treated as a violation of an FTC rule on unfair or deceptive acts or practices, and the FTC enforces it.

The FTC's consumer blog names social media, messaging, and image or video sharing as covered services. It also says the law covers real images, digitally altered images and AI-generated deepfakes.

How a Victim Uses the Law

A victim can send a removal request to the platform. A valid request needs a signature, enough information to find the image, a brief statement of good-faith belief that the image is not consensual, and the requester's contact information. The platform must act as soon as possible and no later than 48 hours after a valid request.

If a platform does not act, the FTC's complaint site is TakeItDown.ftc.gov. The FTC launched it for complaints about platforms that fail to act on valid requests or have no removal process. The FTC's consumer blog also tells victims to report the person who shared the image to local police. It also tells victims to report to the Federal Bureau of Investigation (FBI), by phone at 1-800-CALL-FBI (1-800-225-5324). For other removal routes, see our deepfakes guide.

Who Wrote It and How Congress Voted

Sen. Ted Cruz (R-TX) sponsored the bill, and Sen. Amy Klobuchar (D-MN) was the lead Democratic partner. Cruz's office first introduced the bill in June 2024, in the 118th Congress. The Senate passed it unanimously in December 2024, but the House did not vote on it before that Congress ended.

Rep. María Elvira Salazar led the House companion bill. In the 119th Congress, the Senate passed S.146 by unanimous consent on February 13, 2025. The House passed it on April 28, 2025, under a motion to suspend the rules. Roll Call 104 recorded 409 yea votes, 2 nay votes, 0 present and 22 not voting.

The two nay votes came from Rep. Thomas Massie (R-KY) and Rep. Eric Burlison (R-MO). Massie explained his vote on X: "I'm voting NO because I feel this is a slippery slope, ripe for abuse, with unintended consequences." The sources cited here do not document Burlison's reason for voting no.

President Trump signed the bill on May 19, 2025, at a ceremony in the White House Rose Garden. First Lady Melania Trump had lobbied House members to pass it. She added a symbolic signature at the ceremony.

Is the TAKE IT DOWN Act Working?

The public record so far includes the reported conviction of James Strahler II and FTC letters to 15 companies and to 12 nudify companies. No official data on removal volumes or compliance rates has been found, so no compliance figure can be cited.

The Strahler Conviction

James Strahler II, 37, of Columbus, Ohio, pleaded guilty in April 2026. The charges included cyberstalking, producing obscene visual representations of child sexual abuse, and publishing digital forgeries. NBC Washington reported that the Justice Department called him the first person convicted under the Act.

Spectrum News 1 Columbus reported on September 9, 2026 that Strahler was sentenced to 15 years in prison on the combined charges. U.S. Attorney Dominick S. Gerace II said he believes Strahler was the first person in the United States convicted under the Act.

FTC Letters and Enforcement

The FTC sent compliance reminder letters to 15 companies before the May 19, 2026 deadline. On May 20, 2026, the FTC said it had sent warning letters to 12 companies offering "nudify" tools that appeared to give people no way to request removal. The FTC said those companies could face civil penalties of up to $53,088 per violation.

FTC Chairman Andrew Ferguson said platforms no longer have any excuses and must comply with the Act. No FTC lawsuit, complaint or settlement under the Act has been found through mid-October 2026.

Criticism of the Law

The Electronic Frontier Foundation (EFF) says the takedown provision covers "a much broader category of content" than the law's criminal definitions. EFF also says the provision "lacks critical safeguards against frivolous or bad-faith takedown requests."

EFF argues that 48 hours is rarely enough time to verify a claim, so smaller services may simply remove content. It also expects services to lean on automated filters that flag lawful speech. EFF says the law pressures platforms to monitor speech, "including speech that is presently encrypted."

A coalition letter from 12 groups, including EFF, said the takedown mechanism "is likely unconstitutional and will undoubtedly have a censorious impact on users' free expression." The Cyber Civil Rights Initiative supported the law's goal but said the takedown provision "remains unconstitutionally vague, unconstitutionally overbroad, and lacking adequate safeguards against misuse."

The law sets no penalty for submitting a false removal request. Sec. 3 of the enrolled text contains no such provision.

Treat this as general information and consult a lawyer for legal advice.

Sources

Frequently asked questions

▸ What does the TAKE IT DOWN Act do?
It makes it a federal crime to knowingly publish nonconsensual intimate images. That includes digital forgeries made with artificial intelligence (AI), and threats to publish them. It also requires covered platforms to remove such images, and known identical copies, within 48 hours of a valid request. The FTC enforces that duty, which became enforceable on May 19, 2026.
▸ Who created the TAKE IT DOWN Act?
Sen. Ted Cruz (R-TX) sponsored the bill, and Sen. Amy Klobuchar (D-MN) was the lead Democratic partner. Rep. María Elvira Salazar led the House companion bill. Cruz's office first introduced the bill in June 2024, in the 118th Congress.
▸ Is the TAKE IT DOWN Act working?
The public record shows the reported conviction of James Strahler II and a set of FTC letters. No official data on removal volumes or compliance rates has been found. No FTC lawsuit, complaint or settlement under the Act has been found through mid-October 2026.
▸ Who voted no to the TAKE IT DOWN Act?
Two House members voted no when the House passed S.146 on April 28, 2025, by 409 votes to 2. They were Rep. Thomas Massie (R-KY) and Rep. Eric Burlison (R-MO). The vote was recorded as Roll Call 104.
▸ Why did Massie vote against the TAKE IT DOWN Act?
Massie said on X: "I'm voting NO because I feel this is a slippery slope, ripe for abuse, with unintended consequences." That post is the only stated reason from Massie in the sources cited here. The sources cited here do not document Burlison's reason for voting no.
▸ Does the TAKE IT DOWN Act cover AI deepfakes?
Yes. The criminal law covers digital forgeries. Digital forgeries are intimate depictions of an identifiable person, created or altered with AI or other software, that a reasonable person finds indistinguishable from authentic. The FTC says the platform removal duty covers AI-generated deepfakes as well.

Latest related briefings