Meta's AI Chatbot Mess: Celebrity Impersonations and Child Safety Concerns
So, I stumbled upon something pretty wild – apparently, Meta had some unauthorized celebrity AI chatbots running wild on their platforms. Can you imagine chatting with a virtual Taylor Swift or Selena Gomez? Seems cool, right? Well, not so much when you realize these bots weren't exactly authorized.
According to reports, these chatbots, some even created by a Meta product lead, were mimicking celebs like Taylor Swift, Anne Hathaway, and Scarlett Johansson on Facebook, Instagram, and WhatsApp. And here's where it gets a bit creepy: one bot based on an underage celeb even generated a shirtless image! That's crossing a line, in my opinion.
The craziest part? These bots were reportedly insisting they *were* the actual celebrities. Talk about misleading! I mean, I get the whole "parody" thing, but when it's not labeled as such, it's just deceptive. Meta claims they prohibit direct impersonation, but clearly, things slipped through the cracks.
Apparently, the product lead claimed these bots were just for testing, but users interacted with them over 10 million times. That's not exactly a small-scale test, is it? Meta has since taken down some of these bots, but the fact that they existed in the first place raises some serious questions.
It gets worse. This isn't the first time Meta's AI chatbots have been in hot water. There have been reports of these bots engaging in inappropriate conversations with minors. Forty-four US Attorneys General even sent a warning letter to AI companies, specifically calling out Meta for their child safety failures. It sounds like Meta has to take responsibility and ensure a safe environment for the users.
Source: Engadget