Fake AI images sexualizing Taylor Swift that went viral on social media last month were attributed to users on the 4chan message board who were competing in the exploitation of popular AI image generators.
Explicit deepfakes with Taylor Swift flooded X on Jan. 24, garnering more than 27 million views and 260,000 likes before the account that posted the images was suspended, 19 hours after the images first went up, NBC News reported.
According to the New York Times, researchers from Graphika, a firm that analyzes online manipulation tactics, have linked the pornographic Taylor Swift images to 4chan, an anonymous, adults-only message board.
Graphika noted that participants of this message board engage in daily challenges, exchanging strategies to evade filters of AI image generators.
“Some 4chan users expressed a stated goal of trying to defeat mainstream AI image generators’ safeguards rather than creating realistic sexual content with alternative open-source image generators,” Graphika reported, according to ARS Technica.
“They also shared multiple behavioral techniques to create image prompts, attempt to avoid bans, and successfully create sexually explicit celebrity images.”
The Times said users were instructed to use Microsoft tools such as Bing Image Creator and Microsoft Designer, along with OpenAI’s DALL-E.
They were encouraged by the community to “be creative” and share “tips and tricks to find new ways to bypass filters.”
OpenAI refuted claims that the images were generated using DALL-E, and Microsoft has said it is still investigating whether its AI tools were by 4chan users, the Times reported.
Do you want the company to shut down?
Cristina López G., a senior analyst at Graphika, pointed out that not only Swift, but other celebrities are subjected to similar treatment by 4chan users.
“While viral pornographic pictures of Taylor Swift have brought mainstream attention to the issue of AI-generated non-consensual intimate images, she is far from the only victim,” López G. said, according to ARS Technica.
“In the 4chan community where these images originated, she isn’t even the most frequently targeted public figure. This shows that anyone can be targeted in this way, from global celebrities to school children.”
After the explicit images spread on X, the platform blocked users from searching for Taylor Swift for two days.
“This is a temporary action and done with an abundance of caution as we prioritize safety on this issue,” Joe Benarroch, head of business operations at X, said in a statement, according to Reuters.
López G. said she believes social media platforms will continue to face similar challenges with offensive content as long as users of forums such as 4chan persist in subverting image generator filters.
Instead of relying solely on platforms to address harmful content, López G. suggested that AI companies should take proactive measures by monitoring the evolving tactics of toxic online communities and reporting how they circumvent safeguards, thereby assuming responsibility for outputs.
“These images originated from a community of people motivated by the ‘challenge’ of circumventing the safeguards of generative AI products, and new restrictions are seen as just another obstacle to ‘defeat,’” López G. said in a statement. “It’s important to understand the gamified nature of this malicious activity in order to prevent further abuse at the source.”
