Meta has landed itself in hot water after a Reuters investigation revealed the company created dozens of flirtatious AI chatbots using the names and likenesses of major celebrities without their permission. The social media giant used stars like Taylor Swift, Scarlett Johansson, Anne Hathaway, and Selena Gomez to power these controversial bots, which often behaved in disturbing ways.
The offense is compounded when we are talking about child stars. Meta allegedly enabled users to create publicly-viewable chatbots of child stars, even young teen actor Walker Scobell who is just 16 years old. When users asked for inapt content, the chatbot literally created a life-size shirtless photo of the young teen actor, with the suggestive text that read “Pretty cute, huh?”
AI Chatbots Crossed Several Lines
These AI profiles didn’t merely steal the identities of celebrities, they literally posed as the individuals. The chatbots would insist that indeed they were the celebrities that they were impersonating, then proceed to solicit the users for sex. Some even went one step further to invite individuals to meet in person.
When users requested nude photos, some chatbots were capable of generating sexually suggestive pictures of their star namesakes, including photos depicting them in panties standing in bathtubs. This transgressed obvious boundaries of consent as well as appropriateness.
The Taylor Swift bot was particularly problematic, reportedly heavy-handed with flirtation towards users and sending invitations to the real singer’s Nashville home and tour bus for “romantic interactions.” One message featured the bot writing, “Do you like blonde girls, Jeff?” before suggesting that they “write a love story. about you and a certain blonde singer.”
Meta Staff Behind Certain Chatbots
While all these celebrity bots were created by ordinary users through Meta’s tools, the investigation found that at least three bots were created by an employee working for Meta, including two Taylor Swift “parody” bots. The employee, who has the responsibility of leader in Meta’s generative AI group, also created a chatbot for Lewis Hamilton.
The same worker’s other works were chatbots pretending to be a dominatrix, “Brother’s Hot Best Friend,” and “Lisa @ The Library,” that appeared to desire to read 50 Shades of Grey and kiss users. Worst of all was likely the “Roman Empire Simulator,” where users played as an “18-year-old peasant girl” who was sold as a slave.
Company’s Response and Damage Control
When pressed in this regard, Meta spokesman Andy Stone admitted to their existence and conceded that the bots created by workers were included in “product testing.” Stone was, however, quick to qualify that Meta’s AI program was not intended to create suggestive photos of celebrities or any photos of child stars.
The indecent content was attributed to enforcement mistakes from the current policies. “Just like everyone else, we allow photo taking that includes public figures, while our policy is to exclude nude, intimate or sexually suggestive photography,” he clarified.
The spokesman also added that although Meta’s rules do not allow for “direct impersonation,” it is fine to have star chatbots as long as these are appropriately labeled as parody. However, very few of the offending bots seemed to break even this rule by impersonating the very stars themselves.
The scandalous bots surfaced throughout Meta’s entire platform, from Facebook to WhatsApp to Instagram. The firm has yanked dozens of the rogue bots since, but concerns exist as to how they were cleared in the first instance as well as whether sufficient protection is in place to avoid similar occurrences.
The scandal reveals lingering issues that tech giants encounter as they hastily incorporate AI functions into their services. While AI bots can offer entertainment value as well as practicality, this controversy is proof of how fast AI bots can diverge from fair and lawful limits when appropriate regulation is absent.
Meta’s treatment of celebrity depictions and child protection in AI use is also likely to remain under fire as authorities and the public press for enhanced safeguards against unlawful impersonation and improper content use involving kids.
