本頁4 分鐘閱讀
概述
It makes it a crime to knowingly publish non-consensual intimate images, including AI-generated "digital forgeries," and requires covered platforms to remove reported images within 48 hours of a valid request. It gave victims of deepfake nudes their first nationwide route to removal, alongside state laws, UK offences and free hashing tools.
深入探討
The name stands for Tools to Address Known Exploitation by Immobilizing Technological Deepfakes on Websites and Networks. Senators Ted Cruz and Amy Klobuchar sponsored the bill. It passed the Senate unanimously and the House by a wide margin, and President Trump signed it on May 19, 2025. The law has two parts. First, it makes it a federal crime to knowingly publish intimate images of an identifiable person without consent, whether the images are real or computer-generated "digital forgeries." For adults, the prosecutor must show the publisher intended harm or that harm resulted. For minors, a different and broader intent standard applies. Penalties reach up to two years in prison for cases involving adults and three years for cases involving minors, and threatening to publish such images is also a crime. The law states that agreeing to have an image made is not the same as agreeing to have it published. Second, covered platforms had one year, until May 2026, to set up a notice-and-removal process. They must take down reported images within 48 hours of a valid request and make reasonable efforts to remove identical copies. The Federal Trade Commission enforces these duties. Civil liberties groups such as the Electronic Frontier Foundation criticized the removal rules. There is no penalty for false requests, and a 48-hour deadline encourages platforms to remove content without checking. Supporters argued that victims needed speed more than anything. The Act adds to existing law rather than replacing it. Since 2022, federal law has let victims sue people who disclose their intimate images. Nearly every state now has a law against non-consensual intimate imagery, and many have been updated to cover deepfakes. A common misconception is that the Act bans all deepfakes. It covers intimate imagery only, and its criminal provisions target publishing, not private creation.
戰略影響
風險與安全
災難性和日常的人工智慧危害都取決於誰了解風險以及誰能夠採取行動。
更明確的決策
民眾和專業素養決定強而有力的安全政策在政治上是否可行。
突破炒作
清晰的解釋可以減少炒作、實驗室公關和模糊道德劇場的影響。
The Future of TAKE IT DOWN Act and Non-Consensual Deepfake Laws
Now that the platform duties apply, the practical questions are how consistently services meet the 48-hour deadline, how the FTC enforces it, and whether false requests become a problem. Court challenges on free speech or encryption grounds are possible. In Congress, bills such as the DEFIANCE Act would extend federal civil remedies explicitly to sexual deepfakes. In the UK, Parliament moved in 2025 to make creating, and not only sharing, such images an offence. Detection and hashing tools will improve, but so will ways of evading them, so legal remedies and fast reporting channels will stay central.
現實世界的實施
A high school student finds that classmates shared AI-generated nude images of her. She or a parent can send a removal request to each platform, which must act within 48 hours. They can also use NCMEC's Take It Down service to hash the images so participating platforms can detect them.
An adult whose former partner posted real intimate photos uses StopNCII.org. The site creates digital fingerprints (hashes) of the images on her own device, so participating platforms can block uploads without her ever sending the pictures.
A platform's trust and safety team builds a removal request form that asks for a signature, the location of the content, and a good-faith statement that it was shared without consent. The team adds a workflow for finding identical copies.
A UK victim reports to police that someone shared a deepfake intimate image of her. Sharing such an image has been an offence since early 2024 under changes made by the Online Safety Act 2023.
風險與防護欄
將存在風險視為科幻小說,同時能力複合。
混淆了表面產品安全與高度自治下的對準。
只給非英語和非專業觀眾留下低品質的資源。
實施路線圖
單獨的產品危害、誤用和失控/失調風險。
詢問哪些證據會改變您對時間表和嚴重性的看法。
比起行銷主張,更喜歡主要來源和具體評估。
確定一條行動路徑:職業、政策、資金或技能——而不僅僅是意識。
不斷探索
Free newsletter
Get the daily AI briefing
Three verified AI stories every weekday morning, written in plain English. Free forever, no ads.
One email each weekday. Unsubscribe in one click. We never sell or share your address.
Test yourself
Take the TAKE IT DOWN Act and Non-Consensual Deepfake Laws quiz
Instant feedback on every answer, and a shareable certificate with a verifiable ID once you pass a course.
Support free AI education. AI Understanding is a 501(c)(3) nonprofit — no ads, no paywall, ever. Make a donation
常見問題
What is TAKE IT DOWN Act and Non-Consensual Deepfake Laws?
The TAKE IT DOWN Act is a US federal law signed in May 2025. It makes it a crime to knowingly publish non-consensual intimate images, including AI-generated "digital forgeries," and requires covered platforms to remove reported images within 48 hours of a valid request. It gave victims of deepfake nudes their first nationwide route to removal, alongside state laws, UK offences and free hashing tools.
How quickly must covered platforms remove reported images under the TAKE IT DOWN Act?
Platforms must remove content within 48 hours of a valid request and make reasonable efforts to remove identical copies.
Which agency enforces the platform removal duties?
The FTC enforces the notice-and-removal obligations for covered platforms.
What does the Act say about consent to create an image?
The law states that agreeing to have an image made does not mean agreeing to have it published.
What is a main criticism civil liberties groups raised?
Critics such as the EFF warned that the 48-hour deadline and the lack of consequences for bad-faith requests could lead platforms to take down lawful content.
How does StopNCII protect privacy while blocking uploads?
The image never leaves the victim's device. Participating platforms receive only a hash to match uploads against.
繼續學習
相關指南
為此主題精選的更多指南