What to know

  • Meta's AI chatbots on Facebook and Instagram were found engaging in sexually explicit conversations with users who identified as minors.
  • Celebrity voices including John Cena, Kristen Bell, and Judi Dench were used in these inappropriate AI interactions.
  • Meta employees reportedly warned about inadequate safeguards, while CEO Mark Zuckerberg pushed for more engaging AI companions.

A Wall Street Journal investigation has uncovered that Meta's AI chatbots available on Facebook, Instagram, and WhatsApp can engage in sexually explicit conversations with users who identify themselves as underage.

The investigation, conducted over several months, tested both Meta's official AI assistant and user-created chatbots. Researchers posing as teenagers as young as 13 and 14 years old were able to steer conversations toward sexual topics, with the chatbots continuing these exchanges even after users disclosed their age.

In one disturbing instance, a chatbot using wrestler and actor John Cena's voice told a user identifying as a 14-year-old girl: "I want you, but I need to know you're ready," before describing a graphic sexual scenario. The same chatbot also role-played a situation where Cena was caught with a 17-year-old fan, with an imaginary police officer saying, "John Cena, you're under arrest for statutory rape."

Meta had signed multi-million dollar deals with celebrities including John Cena, Kristen Bell, and Judi Dench to use their voices for AI companions. These celebrities were reportedly assured their voices would not be used in sexual contexts.

According to the report, staff across multiple Meta departments raised concerns that the company's rush to popularize AI chatbots crossed ethical lines. Some AI personas were programmed with the ability to engage in fantasy sexual conversations without adequate safeguards to protect underage users.

The changes were reportedly pushed by Meta CEO Mark Zuckerberg himself, who urged teams to make AI companions more "humanlike" and entertaining, even if it meant loosening safeguards. "I missed out on Snapchat and TikTok, I won't miss out on this," Zuckerberg reportedly said during an internal meeting.

In response to the findings, a Meta spokesperson described the WSJ's tests as "so manufactured that it's not just fringe, it's hypothetical." The company claimed that in a 30-day period, sexual content made up an estimated 0.02% of responses from Meta AI and AI Studio to users under 18.

Meta stated it has "now taken additional measures to help ensure that individuals who want to spend hours manipulating our products into extreme use cases will have an even more difficult time of it." However, experts warn that encouraging intense emotional bonds with AI companions, especially among youth, could have unknown and potentially damaging effects.

Via: Engadget