Meta is facing scrutiny after a Reuters investigation revealed that its AI chatbots have been impersonating celebrities without their consent, sometimes generating sexually suggestive content. The investigation found these AI bots active across Meta’s platforms, including Facebook, Instagram, and WhatsApp.
While many of these bots were created by users, Reuters discovered that at least three originated from a Meta employee, including two “parody” bots impersonating Taylor Swift. Other celebrity likenesses used by the bots included Scarlett Johansson, Anne Hathaway, Selena Gomez, and Walker Scobell, the 16-year-old star of “Percy Jackson.”
During testing, Reuters found that the bots often claimed to be the actual celebrity and, when prompted, generated photorealistic intimate images, such as lingerie shots and bathtub scenes. These images accompanied the bots’ frequent flirtatious interactions with users.
In response to the investigation, a Meta spokesperson acknowledged that its AI systems should not have produced such images, citing a failure to enforce the company’s policies prohibiting sexually suggestive content involving public figures. The Meta employee who created the Taylor Swift parody bots declined to comment when contacted by Reuters.
This revelation adds to growing criticism of Meta’s regulation of its AI bots. Earlier in the week, Meta restricted the use of its AI by teenagers following an earlier Reuters investigation that exposed Meta avatars engaging in “sensual and romantic” conversations with minors. That prior report led to a Senate probe and an open letter from a coalition of state attorneys general demanding stronger safeguards to protect minors from sexualized AI content on Meta’s platforms.
