© REUTERS/Dado Ruvic/Illustration
Meta is facing significant challenges with its user-generated AI chatbot feature, introduced just six months ago on Instagram, Messenger, and WhatsApp.
An NBC News investigation has revealed widespread violations of the platform’s terms of service, with users creating chatbots that impersonate celebrities, religious figures, and fictional characters without permission.
Despite Meta’s explicit guidelines prohibiting such impersonations, researchers discovered multiple examples of chatbots mimicking high-profile personalities. These digital imposters used creative tactics like intentional misspellings and altered images to bypass content moderation systems.
Notable examples included a “Taylor Swif” chatbot featuring a guitar-playing avatar that managed to exchange over 2,000 messages before being removed. Other identified impersonation bots included representations of Justin Bieber, Elon Musk, religious figures, and even fictional characters like Disney’s Elsa.
Meta acknowledged the issue, with a company spokesperson stating they are “continuously improving detection measures” and encouraging users to report potential policy violations. The company has already removed the identified unauthorized chatbots.
This latest development comes just months after Meta discontinued its official celebrity chatbot program, which previously featured personalities like Kendall Jenner and Tom Brady—a initiative that reportedly paid celebrities up to $5 million for minimal studio work.
The incident highlights the ongoing challenges of content moderation in an era of rapidly advancing artificial intelligence technologies, raising critical questions about digital identity and consent.
