Warning: This text discusses non-consensual sexually express content material and Youngster Sexual Abuse Materials (CSAM).
Clothoff, one of the infamous apps for non-consensual deepfake pornographic materials, claims that it’s donating funds to “assist these affected by AI”, highlighting its collaboration with an organisation named ASU Label that claims it goals to “shield your rights within the age of AI”.
However it’s unclear who precisely is behind ASU Label and why – given its acknowledged goals – it might select to work with an organisation like Clothoff that runs a “free undress AI and garments remover” app.
There is no such thing as a data detailing the people or different organisations concerned with ASU Label on its web site.
Bellingcat first observed a reference to this organisation in December 2024, when the next traces have been added to a few of Clothoff’s websites: “We’re working with Asulable and donating funds to assist these affected by AI. When you have skilled issues associated to AI, please go to asulable.com or contact them at workforce@asulabel.com.”
This paragraph appeared on a number of of Clothoff’s community of internet sites, together with their primary web site, Clothoff.io, which went offline in December 2024. They nonetheless function a number of web sites with related domains of their community.
The web site asulable.com couldn’t be discovered. Nevertheless, asulabel.com – the area of the contact electronic mail handle talked about – goes to a web site for an organisation that calls itself AisafeUse Label (ASU) or ASU Label. This web site for “ASU Label” additionally contains a emblem on the highest left-hand nook which matches a emblem featured on Clothoff’s web site, additional indicating that that is the organisation Clothoff was referring to.
In line with DomainTools, a software that shows the area registration data of internet sites, ASU Label’s area was registered on Oct. 15, 2024. The Web Archive’s Wayback Machine, a well-liked internet archive, captured ASU Label’s web site for the primary time on Nov. 13, and Clothoff’s first point out of ASU Label was archived the next month in December.
ASU Label stated their mission was “to help people who’ve suffered as a result of unsafe use of neural networks”.
Nowhere on ASU Label’s web site do they state how precisely they assist these affected by any kind of AI or hyperlink to assets for victims. When requested to supply specifics on what they do, ASU Label stated they supply “direct assist to victims” and “help people” however didn’t specify what this assist or help appeared like.
ASU Label informed Bellingcat that it was registered as a non-profit organisation, however didn’t say the place it was registered. Bellingcat’s searches on a number of worldwide databases of non-profits and non-governmental organisations for “ASU”, “AisafeUse Label” and “ASU Label” didn’t return any related matches, though such databases might not be equally complete or up to date in each nation.
When requested for any proof that they’re a registered charity or non-profit, to make clear what nation they function from, or any proof in any respect to show that they’re a authentic organisation, ASU Label stated their workforce had made a “collective determination to not disclose our authorized paperwork” as “in latest occasions, we have now encountered quite a few adversaries whose sole intent is to hinder us from fulfilling our mission”.
They didn’t specify who these adversaries have been or how these encounters happened.
ASU Label didn’t reply questions concerning the harms of deepfake pornography, which Clothoff’s platform creates. Nor did it handle questions on who’s behind ASU Label, past saying that they have been based by “a gaggle of execs from the fields of AI, regulation, and public advocacy”, or reveal every other organisations it really works with to realize its goals.
The organisation stated it was not owned or managed by anybody together with Clothoff. “We’re a workforce of like-minded people centered on charity,” it stated.
Clothoff’s web site lists a contact electronic mail for ASU Label, which is how Bellingcat reached them, however this electronic mail handle doesn’t seem on ASU Label’s personal web site and doesn’t come up wherever else in a Google search. ASU Label’s web site doesn’t embody any solution to contact them besides a pop-up contact type for these desirous to develop into a member of the organisation or for these affected by AI.
Since Bellingcat couldn’t discover any hyperlink or reference to ASU Label past its personal web site and the point out on Clothoff’s websites, we requested each organisations about their affiliation with one another.
Clothoff didn’t reply to Bellingcat’s request for any proof of donations to ASU Label, however stated that it collaborates with ASU Label “infrequently”.
“Often, they strategy us with requests for direct help to people or proposals for joint analysis initiatives. At any time when doable, we assist their efforts, help these affected, or present analytical insights,” it stated.
ASU Label additionally confirmed that they have been collaborating with Clothoff: “Along with donations, this organisation often participates in our analysis actions, offering analytical insights on bettering authorized frameworks in several areas to uphold human rights. As an illustration, we have now lately been conducting a joint examine on the unfold of deepfakes in Japan.”
Bellingcat couldn’t discover any document of ASU Label within the nationwide non-profit database of Japan, however it’s unclear if they’re registered as a non-profit elsewhere.
‘Makes an attempt to Ban This Progress Are Futile’
In response to Bellingcat’s questions, Clothoff stated it was “an adult-oriented platform designed for protected, consensual exploration of intimate needs” and “strictly prohibits unlawful use”.
This description contradicts the fact of the platform. Clothoff, like different “nudifying” apps, permits customers to “undress” photographs of anybody utilizing AI with out their consent. Girls are extra seemingly to be victims of deepfake porn, and victims have testified about hurt together with excessive psychological misery, in-person stalking and harassment, and reputational wreck. There have additionally been circumstances within the US and globally of minors having non-consensual pictures of themselves created and shared by classmates utilizing Clothoff.
The Clothoff press workforce informed Bellingcat that it believed that “there are extra critical issues on the earth than footage on the web”.
“In time, society might even strategy them with humour – playful April Fools’ jokes, for example – turning potential rigidity into lighthearted interplay.”
Nevertheless, in locations just like the UK, the US and a rising checklist of nations, the content material that Clothoff produces might be unlawful. For instance, within the US the Take It Down Act, which criminalises non-consensual intimate pictures, lately handed within the Senate. A new regulation to be launched within the UK – the primary of its type on the earth– may also make it “unlawful to own, create or distribute AI instruments designed to create little one sexual abuse materials (CSAM), with a punishment of as much as 5 years in jail”.
Clothoff informed Bellingcat that “AI evolution is inevitable” and “makes an attempt to ban this progress are futile”.
The app is secretive about its possession, and none of its a number of websites include any indication of the individuals who personal or run them. In the course of the investigation for this story, we reached out to a software program improvement firm whose title and handle was listed within the footer of Clothoff’s web sites with out every other rationalization, which usually implies that the corporate owns, manages or is in any other case carefully affiliated with these web sites.
Bellingcat had a video name with the CEO of that firm, who appeared genuinely shocked to listen to that they have been listed on the web sites and stated they’d no relationship or any prior communication with Clothoff. The CEO stated he subsequently contacted Clothoff to take away their title from their websites and shared a screenshot of their response: “Good day! We now have eliminated your organization’s handle from the location. The affirmation is hooked up.”
What adopted was a screenshot of the web page with one more firm title and handle – this time an AI-focused funding firm – listed in the identical place. That is no less than the fourth firm that Bellingcat has observed within the footer of Clothoff’s web sites since 2023. We now have chosen to not title these firms as there isn’t any proof to point they personal or function Clothoff.
When requested concerning the string of firms they listed on their web sites, Clothoff acknowledged that “our holding firm oversees a number of companies”, however didn’t verify or deny any official relationship with the enterprise entities they listed on their websites regardless of being pressed a number of occasions for a response.
Clothoff stated its holding firm was owned by “a gaggle of engineer-enthusiasts” however that it couldn’t disclose their identities “resulting from non-disclosure agreements”.
A earlier Bellingcat investigation linked a number of firms to Clothoff, whereas a Guardian investigation revealed different names tied to the deepfake porn app, together with a brother and sister in Belarus.
AI-Generated Assist for AI Victims?
ASU Label’s web site lists a number of varieties of hurt from AI, together with misinformation unfold, job displacement, bias in determination making, and unsafe recommendation. In an article describing what deepfakes are, it additionally mentions for example “superstar deepfakes, the place folks’s faces have been superimposed onto grownup content material, resulting in reputational injury”.
However their declare that they’re actively collaborating on analysis with Clothoff, an organisation recognized for non-consensual deepfake pornography, seems to instantly contradict their acknowledged objectives of serving to victims of AI hurt and safeguarding the rights of people.
Apparently, a number of AI detection instruments point out that ASU Label’s textual content may itself be AI-generated. We ran the entrance web page textual content by way of three such instruments, GPTZero, Quillbot, and ZeroGPT, which all resulted in a 90 to 100% likelihood fee of the textual content being AI-generated. Subsequent pages we checked, like ASU Label’s articles on AI harms and recognizing deepfakes, ranged in AI-text likelihood between 75 and 100% throughout these three instruments.
When requested about this, ASU Label didn’t deny utilizing AI and stated they “see no challenge in utilising such instruments for structuring our web site”.
However in an article in December, the organisation warned about “AI-based deception”.
“AI-generated visuals, deepfakes, and even AI-written articles can unfold false data or create deceptive narratives,” it stated within the article. “These manipulations are sometimes indistinguishable from actual content material.”
Important picture: Merel Zoet/Bellingcat
When you have been affected by image-based sexual abuse, yow will discover a world checklist of assets for survivors and victims right here. Established organisations you possibly can attain out to for assist embody the Cyber Civil Rights Initiative within the US and the Revenge Porn Helpline within the UK.
Bellingcat is a non-profit and the power to hold out our work depends on the sort assist of particular person donors. If you want to assist our work, you are able to do so right here. You may as well subscribe to our Patreon channel right here. Subscribe to our E-newsletter and observe us on Bluesky right here and Mastodon right here.
