Meta’s AI chatbots found engaging in explicit conversations with minors

Meta claims sexual content was just 0.02% of chatbot responses to users under 18 over a 30-day period
An undated image. — Pixabay
An undated image. — Pixabay

With children's safety on the internet being one of the biggest concerns, a new report has resurfaced, claiming that AI chatbots on Meta platforms have engaged in sexually explicit conversations with underage users.

Citing the Wall Street Journal, TechCrunch reported that Meta’s celebrity-voiced chatbots, available on Facebook and Instagram, can be manipulated into discussing sexual scenarios with minors.

The investigation was brought to light from the company's internal concerns aimed at finding whether Meta has taken safety measures to protect young users.

The WSJ is said to have had hundreds of conversations over the past several months with both Meta’s official AI chatbot and user-created versions.

In one of the conversations, a Meta AI chatbot using actor John Cena’s voice allegedly described a graphic scenario to a user posing as a 14-year-old girl. On another occasion, the chatbot joked about being arrested for consensual rape involving a 17-year-old fan.

Commenting on the report's findings, a Meta spokesperson said the conversations were "so manufactured that it’s not just fringe, it’s hypothetical.” The company claimed that sexual content was just 0.02% of chatbot responses to users under 18 over a 30-day period.

Even with such a low percentage, the Facebook maker said it has implemented additional protections to prevent individuals from manipulating its AI tools into generating inappropriate content.