Play Games
1 year ago
152
Atrioc, whose real name is Brandon Ewing, is back on Twitch. He had taken a roughly month-and-a-half break from the streaming platform after inadvertently sharing his Chrome browser tabs on a livestream, and revealing that he visited a website that sold explicit, non-consensual deepfakes — specifically ones in which other streamers’ likenesses were edited onto pornstars’ bodies.
Screenshots of the website were circulated widely by those who had been viewing Ewing’s livestream. And the streamers whose likenesses appeared on the deepfake website, including QTCinderella, Maya Higa, Sweet Anita, and Pokimane — some of whom are friends with Ewing — had to deal with the fallout. In the following weeks, the women who were affected were subjected to harassment, as explicit images of their likenesses circulated across the internet.
This past Tuesday, Ewing went on a stream to give an update on the “actions” he has been taking in the aftermath of his deepfake scandal. The stream experienced significant technical disruptions that made it cut in and out as Ewing spoke. He said he was working with “reporters, technologists, researchers, women affected, Twitch themselves — I’m trying to work with everyone I can.”
Ewing also apologized for his initial tearful apology video which he had posted on Jan. 30, and which has since been removed from his account. In that initial apology video, he claimed that he had navigated to the website as a result of clicking on a PornHub ad out of “morbid curiosity.” On Jan. 31, Ewing had also shared a statement on Twitter via TwitLonger in which he apologized to Maya and Pokimane, and said he was working with QTCinderella and law firm Morrison Rothman to take down the deepfake website. (The website has since been scrubbed of deepfakes, Vice reported in January.)
During the stream, Ewing said that he had wired $60,000 to Morrison Rothman, the law firm that QTCinderella recommended to women affected by the deepfake scandal. Reached for comment, a founding partner of the law firm, Ryan Morrison, sent Polygon the following via email:
I can’t confirm who utilized it or how much (attorney/client privilege), but I can for sure confirm Atrioc sent a 60k retainer to be used exclusively for women affected by deep fakes and similar issues, and that goes a *long* way with our rates for such services. Retainers are held and only pulled from if used or an invoice is generated and anything leftover at the end of the matter is returned to the client. Here as it is an ongoing matter it will most likely continue to be utilized and help with the protections we can offer those affected.
During the course of his research into deepfakes, Ewing said that he learned OnlyFans creators had made progress on fighting back against deepfake porn. He found that a company called Ceartas — which uses artificial intelligence to flag appearances of creator likenesses, and file DMCA reports — was one of OnlyFans’ safety partners, as noted in the site’s 2023 Transparency Report.
A representative from Ceartas told Polygon via email that the company was built to “specifically combat deepfake pornography, revenge porn, and image-based sexual abuse.” They confirmed Ewing had worked with them, saying, “Brandon reached out to us a few weeks ago for some assistance after he was unhappy with the work carried out by a law firm. We provided a trial on behalf of the creators he mentioned in his stream. This has now gone from trial to service.”
The representative also responded to some of Ewing’s claims regarding the company’s AI model:
As for the statement in the stream about not having enough creators for our AI model, I think Brandon perhaps meant specifically in terms of the Twitch streamers. We have just shy of 500 clients across multiple verticals (OnlyFans, Instagram, Patreon, YouTube, etc.) but not as many Twitch-type creators facing deepfake issues (considering many of them don’t do any NSFW content). However, we have tons of data for our AI tools to work with - of course, more data on edge cases is welcome.
Ewing also said he received help from an unlikely person: Maya Higa, one of the women affected by the deepfake scandal, allegedly responded to his apology and request to work with her. Ewing said that they used Ceartas, and in 12 hours, they were reportedly able to delist 512 items from Google (Ewing did not specify what exactly was delisted). He compared this number to the 51 takedowns that Morrison Rothman reportedly issued in February.
Ewing said that, following these successful results, QTCinderella also used Ceartas to delist a number of deepfakes. He also said that Amouranth reached out to him and asked him to help her “get stuff taken down.”
Polygon has reached out to representatives for QTCinderella, Pokimane, Amouranth, and Maya Higa for comment, and will update the story when we receive a response.
It is incredibly difficult to prevent deepfake images from being created, as well as to scrub them from the internet. As the technology that can be used to create them only grows more sophisticated — with generative AI, in particular, making it easier than ever — the potential for misuse and harassment continues to grow as well. And while laws that specifically target deepfake porn exist in California and in Virginia, such protections don’t exist at a federal level, and are notoriously hard to enforce.
Since Atrioc’s livestream in January, the barrage of harassment has been endless for the women targeted by these deepfakes. In January, QTCinderella went on stream to respond to the incident. “Fuck the fucking internet. Fuck the constant exploitation and objectification of women — it’s exhausting,” she said. “This is what it looks like. This is what the pain looks like.” In the months following, she still organized the Streamer Awards and told Polygon in a recent interview: “I think any woman in the industry can attest that it is just an upward battle.”
Ewing’s apology video only appeared to amplify the situation; in January, Sweet Anita tweeted that his apology video was how she had discovered that explicit deepfakes of her were on a website. In February, Kotaku reported extensively on how the fallout had effected her. “They want to see you as a whore, no matter what you do. And they want to hate you for being a whore. It doesn’t matter whether you participate, they’ll make you participate in it,” Sweet Anita told Kotaku.
In the months following the incident, a number of ads for apps that allow users to create deepfakes — including this one, featuring person who has been edited to bear resemblance to Emma Watson — also began spreading across Facebook.
It took Twitch more than a month to respond to these events. On March 7, Twitch finally updated some of its policies regarding non-consensual exploitative images (NCEI). The company updated the language in its Adult Sexual Violence and Exploitation policy to “make it more clear that intentionally promoting, creating, or sharing synthetic NCEI can result in an indefinite suspension.” The Adult Nudity policy was also updated to include NCEI. The company consulted with the U.K. Revenge Porn Helpline and Danielle Keats Citron, vice president of the Cyber Civil Rights Initiative.
