President Trump has now signed Take It, criminalizing sexual deep strikes at the federal level at the United States. Meanwhile, the Civitai community aims to “clean up its behavior” about NSFW AI and celebrity output, ultimately failing to appease payment processors, leaving the site looking for alternatives or facing closures. It’s all in just two weeks since the world’s oldest and largest deep ice porn site offline…
It’s been weeks for unregulated image and video deep production status. Just two weeks ago, the community shared the number one field of celebrity Deepfake porn. Mr. Deepfakes has been a seat on global AI celebrity content after a sudden dominant and high-profile position for more than seven years. By the time it was down, the site received an average of more than 5 million visits per month.
Background, the deep Mr Fakes field in early May; illustration, pause notice, has now been replaced by 404 errors, as the field was apparently purchased by unknown buyers on May 4, 2025 ( Source: mrdeepfakes.com
Mr. Deepfakes’ service cessation was formally attributed to the withdrawal of “critical service providers” (see the illustration image above, which was replaced by domain failures within a week). However, a collaborative news investigation has directly revoked the key figures behind Mr. Diepfakes before the closure, thus causing the website to be closed for the individual and/or legal reasons.
Around the same time, Civitai, a commercial platform widely used in celebrities and NSFW Loras, has taken a series of unusual and controversial self-censorship measures. These influence deep generation, model hosting, and broader new rules and restrictions, including a full ban on certain marginal NSFW fetishes. and what it calls “extremist ideology.”
These measures are facilitated by payment providers apparently threatening to extract services from the field unless changes are made to the description of NSFW content and celebrity AI.
Civitai cuts off
As of today, it seems that the measures taken by Civitai have not yet been added to the visa and Mastercard: New Post† Community Engagement Manager Alasdair Nicoll revealed on the website that Civitai’s card payments (its “buzz” virtual currency system is mainly powered by real-world credit and debit cards), will be stopped from this Friday (May 23, 2025).
This will prevent users from updating monthly memberships or purchasing new buzz. Although Nicoll recommends that users can maintain current membership privileges by switching to annual membership (costs)†† $100-$550 USD) The future of the domain is currently somewhat uncertain (it should be noted that annual membership is online at the same time as the loss payment processor was announced).
Regarding the lack of payment processors, Nicole says “We are talking to every provider who is satisfied with AI innovation”.
As for the recent failure to adequately rethink the website’s policies around Celeb AI and NSFW content, Nicoll said in his post:
“Some payment companies label the tags that generate AI as high risk, especially when we allow mature content (even legal and conditional) generated by users. This policy choice, without anything the user does, forces the cutoff.”
The comment from user “Faeia”, designated as Chief of Staff* in the company’s Civitai Profile*, adds context to this announcement:
“Just to clarify, we were removed from the payment processor because we chose not to delete NSFW and adult content from the platform. We remain committed to supporting a variety of creators and are working on alternative solutions.
As a traditional driver of new technologies, it is not uncommon for NSFW content to start interest in domains, technologies or platforms – simply to allow the initial believers to be denied once they have established enough “legal” capital and/or user funds (i.e., to survive in the NSFW context, and to allow entities to survive enough users.
For some time, Civitai will follow Tumblr and take other plans under this route toward “disinfection” products, ready to forget its roots. But the additional and growing controversy/stigma surrounding AI-generated content Any form In this case, it represents a cumulative weight that seems to prevent last-minute rescue. At the same time, the official announcement recommends that users use encryption as an alternative payment method.
Fake out
The appearance of President Donald Trump enthusiastically signed the Federal Shootdown Act, which could affect some of these events. The new law criminalizes the distribution of non-consensual intimate images, including deep strikes generated by AI.
The legislation stipulates that the platform deletes marked content within 48 hours and is supervised by the Federal Trade Commission. The criminal provisions of the law come into effect immediately, allowing individuals who intend to publish or threaten to publish involuntary intimate images (including deep cakes generated by AI) within the United States.
While the law has received rare bipartisan support and is supported by tech companies and advocacy groups, critics argue that it could curb legitimate content and threaten privacy tools like encryption. Last month, the Electronic Boundary Foundation (EFF) announced opposition to the bill, asserting that the goal of its authorized revocation mechanism is to be a broader substance than the narrow definition of involuntary intimacy images found elsewhere in the legislation.
“The removed clause applies to a wider range of content categories (any image involving intimate or sexual content), rather than the NCII definition of other locations in the bill. The revocation also lacks critical safeguards for requirements of rash or undesirable beliefs.
Services will rely on automatic filters, which are notoriously blunt tools. They often tag legal content, from fair use of comments to news reports. The law’s pressing time requires applications and websites to remove voice within 48 hours, and there is little time to verify that the speech is really illegal.
“As a result, online service providers, especially smaller service providers, may choose to avoid heavy legal risks by simply giving a speech rather than trying to verify it.”
Now, it has been a year since the law enacts bill to create formal notices and fishing processes, allowing affected individuals or their representatives to invoke the statute to seek to remove content.
This means that, despite the immediate effect of the criminal provisions, legally, the platform has no obligation to comply with the revocation of the infrastructure (such as receiving and processing requests) until the window of one year has passed.
Does the act of putting it down cover the celebrity content generated by AI?
Although the Knockdown Act crosses all national boundaries, it is not necessarily an AI-driven media for all celebrities. This act criminalizes the distribution of involuntary intimate images, include AI-generated deep strikes only if the person being portrayed has one Reasonable privacy expectations:
The law states:
“(2) Crime involving real intimate visual descriptions.
“(a) Participate in adults [for evidentiary, reporting purposes, etc.]for anyone, in interstate or foreign business, will be illegal
“(i) Intimate visual descriptions are obtained or created when a person knows or reasonably knows a identifiable individual’s reasonable expectation of privacy expectations;
“(ii) the description is not voluntarily exposed by an identifiable individual in a public or commercial setting [i.e., self-published porn];
“(iii) depicts not issues of public concern;
“(iv) Post an intimate visual description –
“(i) is intended to cause harm; or
“(ii) Inflict harm on identifiable individuals, including psychological, financial or reputational injuries.
The “reasonable expectation of privacy” adopted here traditionally does not support the rights of celebrities. According to the final case law, it is even possible Explicit AI-generated content involving public numbers in public or commercial environments may not fall under the ban of the bill.
Regarding the final terms of determination degree Legally speaking, harm is famous for flexibility, and in this sense there is no particular novelty to the legislative burden. but, intention The harm caused appears to limit the scope of the behavior to the context of “revenge porn” in which (unknown) former partners published real or fake media content from other former partners.
While the “hazard” requirement of the law does not seem to be suitable for anonymous users to post descriptions of AI-generated celebrities, it may be more relevant in stalking schemes, in which case a wider pattern of harassment supports the conclusion that a person intentionally and maliciously targets public figures on multiple fronts.
Although the bill’s reference to “covered platforms” does not include private channels, such as signals or emails in revocation clauses, in addition, it is only applicable to obligations to implement a formal evacuation mechanism by May 2026. This does not mean that involuntary AI or real DEPTIONS shared through private communications does not exceed the scope of legal criminal offences.
Obviously, the lack of a site reporting mechanism does not hinder the affected parties from reporting the current illegal content to the police; none of these parties exclude any conventional contact information that may complain and require the removal of the problematic material.
The right to leave
Public and media criticism of Deepfake content for more than seven years seems to have reached its highest level in a very short period of time. But while Take It Down Act provides a thorough federal ban, it may not apply in every case involving AI-generated simulations, and in the case of a patchwork of growing state-level deep bill legislation, the law often reflects “local interests.”
In California, for example, the California Celebrity Rights Act restricts the exclusive use of celebrity identities, and even after their death, limits themselves and their own property. Instead, Tennessee’s Elvis Act focuses on protecting musicians from unauthorized AI-generated replicas of sound and imagery, each reflecting the targeted approach of targeting groups that stand out at the state level.
Now, most states have targeted deep-strike laws, although many cannot clarify whether these protections extend to private and public figures. Meanwhile, in some cases, it is reported that in some cases it has been reported that it will help inspire Donald Trump’s support for the new federal law may be contrary to constitutional barriers.
† Archived version: https://web.archive.org/web/20250520024834/https://civitai.com/articles/14945
†† Archived version (excluding monthly price): https://web.archive.org/web/20250425020325/https://civitai.green/pricing
* The actual “Chief of Staff” of the CEO of Civitai is listed under an unrelated name, while the similar-sounding “Faiona” is the official Civitai employee moderator for the domain Subreddit.
First published on Tuesday, May 20, 2025
