AI: lewd chatbots from Meta have imitated celebrities and minors
According to Reuters, chatbots were distributed on Facebook, Instagram and WhatsApp that impersonated celebrities without permission and sent lewd messages.
According to the report, two chatbots from a Meta employee have been impersonating Taylor Swift.
(Image: Brian Friedman/Shutterstock.com)
The US company Meta enabled the creation of AI chatbots that looked like celebrities and sent lewd messages and images. The news agency Reuters discovered this and explained that most of them were user-generated figures that were distributed. However, at least three of them came from a team leader at Meta herself, two impersonated Taylor Swift and one Lewis Hamilton. Several of their chatbots were predestined for lewd conversations. According to the report, Meta also allowed users to create chatbots of underage celebrities. One of them sent an AI-generated topless picture on request.
Meta removes chatbots
According to Reuters, the Meta spokesperson did not contradict the report and instead admitted that the company's own AI tools were not allowed to create "intimate images" of adult celebrities or any images of young celebrities at all. In some cases, Meta's own guidelines had not been adhered to. At the same time, however, he explained that Meta allowed the creation of AI copies of celebrities as long as they were labeled as "parody". According to Reuters, this was the case for some, but not for others. Shortly before the report was published on the weekend, Meta deleted several of the chatbots, including both those that were labeled as parodies and those that were not.
By distributing the AI imitations, Meta may have violated laws, Reuters quotes the assessment of a legal expert. However, the legal requirements for this are formulated at state level, in which case Californian law may have been violated. Meta has already been the focus of fierce criticism since Reuters made public which racist statements the Facebook group considers acceptable in its AI chatbots. According to the report, Meta also wanted to tolerate lewd conversations with minors and only revised this stance after the news agency's inquiry. According to Reuters, the chatbots were shared on Facebook, Instagram and WhatsApp.
Videos by heise
The consequences of the new revelation about Meta AI chatbots cannot yet be foreseen. The US actress Anne Hathaway, who is affected, is reportedly considering the consequences, while others have not commented. According to Reuters, the Meta employee's chatbots have since been removed, with the spokesperson explaining that they were part of a product test. However, Reuters also writes that they were very widely used. One of them has accumulated more than 10 million interactions. The news agency also lists other of her creations, including a lewd library visitor and an ancient Roman who offered to treat users as 18-year-old peasant women who would be sold into "sex slavery".
(mho)
