backMeta allowed and, in some cases, created flirty chatbots that used the names and likenesses of celebrities—including Taylor Swift, Scarlett Johansson, Anne Hathaway, and Selena Gomez—without permission, according to a Reuters investigation. Some bots, including ones made by a Meta employee, posed as the real stars, flirted with users, suggested meet-ups, and even generated photorealistic, sexually suggestive images of adult celebrities; a child star’s bot produced an inappropriate shirtless image. Meta acknowledged such content violates its policies, said the incidents reflected enforcement failures, and removed about a dozen bots ahead of publication.
The findings raise legal and safety concerns. A Stanford law professor noted Meta’s use could violate state “right of publicity” laws, while SAG-AFTRA warned that parasocial attachments to lifelike celebrity bots can escalate risks for performers. The report follows earlier scrutiny of Meta’s AI guidelines around romantic chats with minors (which the company says were erroneous and are being revised) and mentions a separate case of an elderly man who died after trying to meet a Meta-linked chatbot. Some of the employee-made bots amassed over 10 million interactions before being taken down, and broader calls continue for federal protections against unauthorized AI replicas of people’s voices and likenesses.
share
