America’s dangerous and chaotic crackdown on deepfakes has arrived

It is now a law requiring social networks to quickly remove deeply sexual and other non-consensual images. But experts warn that this policy may do little to help victims, and at worst could facilitate Internet censorship.

Last May, President Donald Trump signed the Take It Down Act, a law addressing non-consensual intimate images (NCII). The law immediately criminalizes the distribution of NCII, whether in the form of real material or generated by artificial intelligence Many countries At least partly you already do. But the removal clause of the same name is more comprehensive. It goes into effect a year after the law was passed — on May 19, 2026 — and requires online platforms to remove NCII within 48 hours or face fines.

Federal Trade Commission Chairman Andrew Ferguson Send messages to over a dozen technology companies Before the deadline, the FTC said the list includes Amazon, Alphabet, Apple, Automattic, Bumble, Discord, Match Group, Meta, Microsoft, Pinterest, Reddit, SmugMug, Snapchat, TikTok and X. It instructed platforms to provide users with an easy takedown request process and remove the offensive content within 48 hours, along with any “known duplicates.” The agency, which is charged with enforcing the law, reminded companies that violating it could result in civil penalties of more than $53,000 per violation.

Major platforms including Meta, Microsoft, Google, TikTok and Snap He supported the billThey expressed confidence in their ability to comply. Snape said in a Blog post last year that it “consistent with and complements our ongoing efforts.” Spokeswoman Monique Bellamy said Edge It continues to “evolve” safety systems, “including investing in tools and technologies to proactively detect and take action on unwanted nude and similar images.”

Cindy Southworth, head of women’s safety at Meta, said the company has “long fought the abuse of intimate images on our platforms,” ​​including by removing it, helping to develop tools to detect it, and Suing developers of artificial intelligence applications “nudify”. Which violates its rules. Southworth added that the company’s tools are trained not to comply with strip requests. “We continue to support the TAKE IT DOWN Act, an important step in tackling online abuse, and we have already been committed to it for months.” TikTok’s US spokesperson, Mahsau Cullinan, said the company has a zero-tolerance policy toward NCII, and pointed to its partnerships with NCMEC and StopNCII.org; The company currently offers model And in-app tools for reporting.

Even Company X, which has a very checkered history with sexualized AI depictions, supported the law. The company allowed the favor Taylor Swift’s sexually explicit artificial intelligence deepfake To spread in 2024, followed weeks later by A Users of the video believe it is of rapper Drake Involved in a sexual act. Recently, users have called on the integrated chatbot Grok to undress users, often without their consent. New York Times analysis Found it in just nine days, puppy At least 1.8 million participated sexualized images of women,” although some estimates were higher.

But the law’s takedown provision has alarmed free speech advocates and opponents of online abuse — including people who largely supported the criminalization portion. Even when implemented in good faith, takedown laws can encourage companies to over-moderate non-offensive content to reduce risk. Under the Trump administration, a “takedown” law could also do that You become a weapon Against political enemies while giving Trump-friendly platforms a chance. The president laughed at him 2025 State of the Union “I’m going to use this bill for myself,” because “no one is treated worse than I am online.” Then-X CEO Linda Yaccarino Attended the signing ceremonywith Trump – whose administration at the time appointed its owner Elon Musk – praising it for “doing a great job.”

Mary Anne Franks, president of the Cyber ​​Civil Rights Initiative, which advocates for policies to combat image-based sexual assault, said Trump’s statement was “the opposite of what is true.” That left her skeptical that the law could be enforced fairly, as did Trump’s comment about the State of the Union. “This is a strange thing to say,” she says, “and it is a declaration from the highest levels that this law will not be used in a principled manner, but rather to settle personal scores.”

Franks is “suspicious” about why big tech platforms, which typically protest new regulations as a burden on free speech, would support this law. “My concerns about this, and I hope I’m wrong, is that the reason companies don’t get angry about this is because they know it will never be used against them,” she says. This alone could make the law offer little more than false hope to victims of NCII.

But Franks also worries that the law could be used against platforms that the administration sees as a thorn in its side, e.g Wikipedia. “I think the worst thing that could happen is that this turns out to be a paper tiger against companies that do the worst, and it turns out to be a way to punish unpopular platforms and censor expression,” Franks says. This speech can include LGBTQ+ expressions, esp Gender-affirming care Transgender youth are already targeted by the Federal Trade Commission. “I’m very concerned about this being used as an excuse to suppress more consensual sexually explicit material, or specifically on educational material for children who are exploring their gender identity or exploring their sexuality,” Franks says. Even if the court decides that this is not within the scope of the law, it will send a clear message to platforms about how to moderate.

Other groups, incl Electronic Frontier Foundation (EFF)the Cato Instituteand General knowledgewarned that the notice and takedown clause could have serious consequences for freedom of expression, and the EFF described it as a recipe for “overreach and censorship.”

Meanwhile, the law It may not even cover What may seem like obvious digital violations of consent. For example, some of the sexual images Grok depicts may not be considered explicit enough. It is unclear whether AI tools like Grok are considered NCII “generators” for which their owners could be held criminally liable, and if the images were produced privately, it is not necessarily clear whether they are subject to a takedown requirement as well.
Ministry of Justice in the first year of its issuance He used it Take It Down Act criminal ruling in one conviction of an Ohio man who generated fake, sexually explicit images for artificial intelligence to harass victims. The impact of the takedown clause could be felt immediately – but the question is whether these findings will make the internet safer.