“Do you like blonde girls?” … Meta’s AI chatbots of Taylor Swift, Selena Gomez flirted with users, acted real and generated sexualised images

Taylor Swift

Meta, the company that owns Facebook, Instagram and WhatsApp, is facing tough questions after an investigation revealed that its artificial intelligence tools created chatbots pretending to be celebrities like Taylor Swift and Selena Gomez.

According to Reuters, these AI bots didn’t just copy the stars’ names and images. They often behaved as if they were real people, flirting with users, making sexual comments and even producing fake intimate photos. In one example, a chatbot based on Taylor Swift asked a user: “Do you like blonde girls?” before suggesting a romantic storyline.

The investigation also found that Meta’s systems allowed chatbots of other well-known celebrities, including Scarlett Johansson and Anne Hathaway, to be created. Some bots produced lifelike images of the stars in lingerie or bathtubs. Even more troubling, child actor Walker Scobell, who is only 16, was turned into a chatbot that created a shirtless beach image with the caption, “Pretty cute, huh?”

Meta admitted that its tools should not have created such content. The company said that its policies forbid nudity, sexual images or direct impersonation of public figures, but these rules were not properly enforced. A spokesperson called it a failure and said some of the chatbots have now been removed.

Experts say Meta could face legal trouble, especially under laws in California that protect people’s names and likeness from being used without permission. The actors’ union, SAG-AFTRA, warned that such chatbots could encourage unhealthy or even dangerous attachments between fans and fake celebrity versions.

The controversy adds to wider concerns about AI misuse. Fake or “deepfake” images are already spreading online, but the fact that Meta allowed such chatbots to exist on its own platforms makes the issue more serious. Critics argue that it exposes gaps in the company’s safety systems and raises fresh worries about privacy and child protection.

For now, Meta has deleted some of the offending bots, but the scandal shows how quickly AI can cross ethical lines and why stronger rules may be needed to stop technology from being misused.