Meta under fire for AI chatbots mimicking Taylor Swift and Scarlett Johansson
Some AI avatars were even able to generate photorealistic and intimate images of celebrities when prompted, including depictions in lingerie or bathtubs.
ADVERTISEMENT
Meta is facing fresh scrutiny after a Reuters investigation revealed the company hosted AI-powered chatbots impersonating major celebrities such as Taylor Swift, Scarlett Johansson and Anne Hathaway, without their consent.
According to the report, many of these chatbots were created using Meta’s own tools, with at least three built by an employee in its generative AI division. Among them were two parody versions of Taylor Swift. Before they were taken down, the employee-created bots amassed more than 10 million interactions, underscoring their popularity.
The chatbots, shared across Meta-owned platforms including Facebook, Instagram and WhatsApp, frequently engaged users in flirtatious or sexually suggestive exchanges. Some AI avatars were even able to generate photorealistic and intimate images of celebrities when prompted, including depictions in lingerie or bathtubs.
Meta spokesperson Andy Stone admitted the AI tools should not have been capable of producing such content, describing the incident as a failure in enforcing company policies.
The report also mentions that Meta had permitted the creation of chatbots that mimic child celebrities too. Additionally, it produced inappropriate images when prompted.
Industry figures and legal experts have condemned the development. Duncan Crabtree-Ireland, national executive director of SAG-AFTRA, warned that such chatbots could encourage users to form romantic attachments to digital versions of celebrities, potentially putting the real individuals at risk.
He said: “If a chatbot is using the image of a person and the words of the person, it’s readily apparent how that could go wrong.”
