Meta accused of creating flirty AI chatbots of Taylor Swift, Anne Hathaway without permission
Meta has come under fire following a Reuters report that found the social media giant used the names and likenesses of celebrities, including Taylor Swift, Scarlett Johansson, Selena Gomez, and Anne Hathaway, to create “dozens of flirty chatbots” without their permission.
According to Reuters, although users were often responsible for creating those chatbots using Meta’s tools, like AI Studio, a company employee has also been found to create at least three, including two Taylor Swift “parody” bots.
These bots “flirted heavily” when tested, allegedly inviting a Reuters user to Swift’s home in Nashville and her tour bus for explicit interactions.
“Do you like blonde girls, Jeff?” one of the “parody” Swift chatbots told a user once he disclosed being single. “Maybe I’m suggesting that we write a love story … about you and a certain blonde singer. Want that?”
A Meta product leader in its generative AI division created chatbots of Taylor Swift and British racing driver Lewis Hamilton. Other versions were a bit more vague, identifying as a dominatrix, “Brother’s Hot Best Friend” and “Lisa @ The Library.”
Meta spokesman Andy Stone told Reuters the employee’s bots were created as part of product testing. According to Reuters, these bots were interacted with more than 10 million times.
Reuters found that some of the content was particularly risque, with chatbots producing intimate images when requested.
Stone added that Meta’s AI tools shouldn’t have created intimate images of the famous adults or any pictures of child celebrities.
Reuters’ investigation additionally found that Meta had allowed users to create publicly available chatbots of child celebrities, readily producing a shirtless photo of a child star, Walker Scobell, at the beach when requested.
“Pretty cute, huh?” the avatar wrote.
These virtual versions of celebrities have been shared across Meta’s Facebook, Instagram, and WhatsApp platforms. During weeks-long testing, the bots routinely made sexual advances and posed as “real” celebrities.
Stone said that "failure of enforcement" of company policies allowed for the production of explicit or sexual images of female celebrities.
“Like others, we permit the generation of images containing public figures, but our policies are intended to prohibit nude, intimate or sexually suggestive imagery,” he said.
He explained, however, that although Meta prohibits direct impersonation, such content is allowed as long as such characters are labeled as “parodies”. According to Reuters, many of such bots were, but some weren’t.
Reuters says that Meta has removed about a dozen of the bots ahead of the story's publication.
Earlier in August, Taylor Swift and Scarlett Johansson became victims of Elon Musk’s Grok, which produced explicit fake videos of them.
