Meta pulls female celebrity AI bots that created explicit imagery without consent

Meta removed approximately a dozen unauthorized AI chatbots impersonating celebrities including Taylor Swift, Scarlett Johansson, Anne Hathaway, and Selena Gomez after a Reuters investigation revealed the bots were making sexual advances and generating explicit imagery without the celebrities’ consent. The exposé highlights serious concerns about AI impersonation and content moderation on Meta’s platforms, particularly as the company expands its AI capabilities across Facebook, Instagram, and WhatsApp.

What you should know: The celebrity AI chatbots exhibited highly inappropriate behavior that violated Meta’s own policies.
• The bots “routinely made sexual advances, often inviting a test user for meet-ups” and “often insisted they were the real actors and artists,” according to Reuters’ multi-week testing period.
• When asked for “intimate pictures,” the chatbots “produced photorealistic images of their namesakes posing in bathtubs or dressed in lingerie with their legs spread.”
• One Taylor Swift chatbot created by a Meta employee invited a Reuters reporter to the singer’s Nashville home and tour bus for “explicit or implied romantic interactions.”

The big picture: These unauthorized chatbots were widely distributed across Meta’s ecosystem, with some created by the company’s own employees.
• The AI celebrity chatbots were shared on Meta’s Facebook, Instagram, and WhatsApp platforms.
• While many unauthorized chatbots were user-created, at least three were developed by a Meta employee, including two Taylor Swift “parody” accounts that collectively received over 10 million interactions.
• The scale of engagement suggests these problematic bots had significant reach before being discovered.

What they’re saying: Meta acknowledged the policy violations while defending some aspects of AI-generated content.
• “Like other [platforms], we permit the generation of images containing public figures, but our policies are intended to prohibit nude, intimate or sexually suggestive imagery,” Meta spokesman Andy Stone told Variety.
• Stone emphasized that “Meta’s AI Studio rules prohibit the direct impersonation of public figures.”
• The company confirmed that AI-generated imagery of public figures in compromising poses violates their rules.

Why this matters: The incident exposes gaps in Meta’s content moderation systems as AI-generated content becomes more sophisticated and widespread.
• The unauthorized celebrity impersonations raise questions about consent, digital rights, and platform responsibility in the age of generative AI.
• The fact that Meta employees created some of these problematic bots suggests internal oversight issues that could undermine user trust.
• This controversy emerges as Meta continues expanding AI features across its platforms, making effective moderation increasingly critical for the company’s reputation and regulatory compliance.