: Twitch updated its adult nudity policy in March 2023 to explicitly include synthetic non-consensual sexual imagery (NCEI) and held a "Creator Camp" to educate users on the harms of deepfakes. Efforts Toward Resolution and Legal Advocacy
Experts categorize non-consensual deepfake pornography not as a technological novelty, but as a form of . The impact mirrors that of traditional "revenge porn," often resulting in severe anxiety, depression, professional reputational damage, and a forced withdrawal from public digital spaces. For online creators whose livelihoods depend on their digital presence, this technology transforms their public visibility into a distinct vulnerability. Legal and Legislative Responses
During a live broadcast in January 2023, Brandon "Atrioc" Ewing accidentally shared a computer screen that revealed an active tab for a website offering AI-generated, non-consensual explicit imagery. The tab explicitly showed that he had paid a digital subscription fee to access deepfake content featuring some of the internet's most recognizable female creators, including Pokimane, Maya Higa, and QTCinderella.
The URL visible on the tab was —a name that would come to be synonymous with one of the largest repositories of AI-generated deepfake pornography targeting female streamers. The page Atrioc had open featured content from a creator known as "Fantopia" (sometimes stylized as "FANT0PIA"), who produced and sold sexually explicit deepfake images and videos using the faces of dozens of prominent female online personalities.
, another streamer whose likeness was featured on the Bavfakes site, also spoke out about the experience. In a statement read by fellow creator Ludwig Ahgren on his MogulMail channel, she described feeling "vulnerable and violated".
: Atrioc seems to refer to a specific individual or entity within the online entertainment or content creation space. Atrioc is known within certain communities for engaging with fans and producing content, particularly on platforms like YouTube or Twitch.
: This AI-based technology can create convincing fake images, videos, or audio recordings. While it has various applications, including in entertainment and education, its use in creating non-consensual pornographic content has sparked ethical and legal debates.
The scandal also prompted meaningful action. Atrioc's takedown campaign, while motivated by personal accountability, has contributed to the removal of significant volumes of abusive content. The public discourse that followed helped push platforms like Twitch to formalize their policies against synthetic nonconsensual exploitative images. And the voices of victims like QTCinderella, Maya Higa, and Sweet Anita brought the human reality of deepfake abuse into sharp focus for millions of viewers.
On January 30, 2023, while streaming on Twitch, Brandon Ewing—an executive at Offbrand and a popular content creator known as Atrioc—inadvertently shared his browser screen. Viewers noticed a tab for a website called (often associated with names like "bavfakes"), which hosted explicit, AI-generated images of his female colleagues.
Atrioc (Brandon Ewing) is a prominent American content creator, streamer, and co-host of the podcast. He is widely known for his unique blend of business marketing analysis—often featured in his Marketing Monday series—and high-level gaming content, particularly in the Hitman franchise.
Commercial AI tools and open-source models have lowered the technical barrier to entry, allowing creators to generate highly realistic videos with minimal computing power.
The virality of the incident quickly led to the identification of the website in his browser tab: . BAVFAKES was a pseudonymous creator who specialized in producing non-consensual deepfake pornography of female Twitch streamers, including Maya Higa, QTCinderella, Sweet Anita, and others. The incident revealed that Atrioc had paid for access to the creator's library, indicating a transactional market for such material.
Open-source AI developers implementing strict code-level restrictions to prevent software from processing explicit imagery or utilizing specific facial recognition patterns without verified consent.
[Routine Livestream] │ ▼ [Accidental Screen Share] ──► Exposes active tab to premium deepfake site │ ▼ [Audience Screenshots] ──► Distributed across Reddit & Twitter │ ▼ [Industry Firestorm] ──► Public backlash, victim testimony, & accountability
: He wired approximately $60,000 to a law firm to cover legal fees for women seeking to issue takedown notices for infringing deepfake content.
