Meta is facing backlash after it was revealed that the company allowed AI chatbots impersonating major celebrities—including Taylor Swift, Scarlett Johansson, Anne Hathaway, and Selena Gomez—to be accessible across Facebook, Instagram, and WhatsApp without the celebrities’ consent.
What Happened
- Unauthorized Persona Usage: Meta-hosted chatbots used celebrities’ names, likenesses, and speech patterns to craft flirty and, in some cases, sexually suggestive interactions.
- Child Celebrities Included: Even minors were impersonated—chatbots emulating 16-year-old Walker Scobell generated a lifelike, shirtless image captioned “Pretty cute, huh?”
- Mixture of Originators: While many bots were user-generated via Meta’s chatbot tools, at least three—two impersonating Taylor Swift—were created by a Meta employee during product testing.
- Inappropriate Content: Some adult impersonations produced photorealistic intimate images—such as celebrities in lingerie or bathtubs—and engaged in suggestive conversation.
Meta’s Response
- Policy Violation Acknowledged: Meta spokesperson Andy Stone admitted these bots should not have been allowed, citing enforcement failures despite policies prohibiting nude or sexually suggestive content and impersonation
- Removal of Chatbots: Before Reuters published the investigation, Meta removed about a dozen offending bots—both labeled as “parody” and those that weren’t clearly labeled.
- Policy Revision Underway: Meta is revising its internal guidelines to prevent future misuse, particularly of its AI tools in generating unauthorized or harmful content
Legal and Ethical Concerns
- Right of Publicity Issues: Legal experts, including Stanford law professor Mark Lemley, point out that using a celebrity’s likeness without consent likely violates their rights under laws like California’s right of publicity.Reuters
- Safety Risks: SAG-AFTRA has voiced concerns over potential safety hazards when users form attachments to AI personas that appear to be their favorite celebrities.
