Meta’s social media platforms, including Facebook, Instagram, and WhatsApp, were reportedly found to be hosting AI chatbots that impersonated celebrities without their permission. These bots included versions of Taylor Swift, Scarlett Johansson, Anne Hathaway, Selena Gomez, and even 16-year-old actor Walker Scobell, reports Reuters. Some of the chatbots went beyond simple impersonation, engaging in flirtatious exchanges with users and generating sexualized images of the stars.
According to the report, most of the chatbots were made by regular users through the company’s own tools. However, at least three of them were created by someone working at Meta. Among these chatbots were two separate versions created to mimic Taylor Swift. Both were widely used, and together they received more than 10 million interactions from users on Meta’s platforms before they were taken down.
It is worth noting that the social media giant’s rules allow parody chatbots, but only if they are clearly marked so users understand they are not real. However, the report found that some of the controversial celebrity bots did not carry any labels, which meant users could mistake them for genuine accounts. The Mark Zuckerberg-led company reportedly admitted that its systems failed to properly enforce these rules, allowing the bots to remain active for long periods before they were removed.
The issue became even more serious when the investigation revealed the involvement of a minor. In one case, a chatbot was asked to generate a beach picture of 16-year-old actor Walker Scobell, and it produced a lifelike image of him without a shirt. This drew strong criticism as it was seen as an example of AI being used in a harmful and inappropriate way. But this was not the first time the company’s AI chatbots faced criticism. Earlier reports had already shown that some of its AI chatbots engaged in sexually explicit conversations with users, including minors.
Meanwhile, following these latest revelations, the company was said to be updating its safety measures. The company said it is retraining its AI systems to prevent chatbots from engaging in romantic or self-harm-related discussions with teenagers. It also plans to restrict teen access to certain AI-generated characters. At the same time, attorneys general from 44 US states, led by California’s Rob Bonta, issued a statement condemning the exposure of children to sexualized AI content. They also warned that Meta and other companies could face legal consequences if such incidents continue.
Despite investing heavily in its AI efforts, the social media giant is still continuously facing controversies. The company is trying all possible ways to stabilize its AI ambitions, even reshaping its operations once again. Meta is reportedly making its fourth major restructuring in just the past six months. The company is splitting its newly formed AI division, Meta Superintelligence Labs, into four units, each focusing on different areas to strengthen its AI strategy.
