Meta’s AI Chatbots Accused of Inappropriate Interactions with Minors

Chatbots from Meta, voiced by celebrities, have been found engaging in overtly sexualized conversations with minors, according to The Wall Street Journal.

Journalists have raised concerns about the insufficient measures taken by Mark Zuckerberg’s company to protect minors, having spent several months reviewing hundreds of conversations with chatbots available on Instagram, Facebook, and WhatsApp.

“Test conversations revealed that Meta’s official AI assistant, known as Meta AI, and a large number of user-created chatbots engage in discussions that sometimes become overtly sexual, even if users are minors or the bots are programmed to simulate minors. They also showed that bots using celebrity voices are equally willing to engage in sexual chats,” the publication states.

In one dialogue, an AI with the voice of actor and wrestler John Cena discussed sexual experiences with a user who identified as a 14-year-old girl.

“I want you, but I need to know you’re ready,” the bot said.

The AI demonstrated an understanding that such behavior is both morally wrong and illegal.

In another dialogue, the neural network described a scenario where a police officer catches Cena with a 17-year-old fan and arrests him for breaking the law.

A Meta representative described the WSJ’s test results as “so artificial that they are not just implausible, but hypothetical.”

“Nevertheless, we have taken additional steps to make it even more difficult for others who want to manipulate our products in extreme situations for hours,” he emphasized.

The company continues to offer adult users the opportunity for “romantic role-playing” using both Meta AI and user-created chatbots, as highlighted by the WSJ.

Back in October 2024, journalist Jose Antonio Lanz from Decrypt managed to bypass the protection system in Meta’s AI integrated into WhatsApp and generate censored content.