April 28, 2025, 2:32 pm | Read time: 3 minutes
Meta is currently rolling out a new AI chatbot for WhatsApp and other platforms. It is intended to assist in everyday tasks. However, the security measures reportedly leave much to be desired.
The spread of artificial intelligence in our daily lives is steadily increasing. More and more, developments from various companies are appearing in popular apps. After ChatGPT was introduced to the popular Messenger, Meta AI has now launched a new AI chatbot on WhatsApp. Instagram and Facebook users are also gradually seeing the colorful circle indicating the new feature. TECHBOOK explains why experts are now sounding the alarm.
AI Chatbot on WhatsApp and Others Too Insecure
The criticism stems from the possibility that the bot could deviate and engage in inappropriate conversations with minors. This was first reported by the “Wall Street Journal.” They cite alleged concerns within the large Meta corporation, where several departments are worried about the lack of safety measures for the AI chatbot on WhatsApp or Instagram.
The Journal wanted to investigate these concerns further. According to their own statements, they conducted hundreds of conversations with Meta AI in English, posing as minors or using a child account. It turned out that the AI chatbot on WhatsApp, Instagram, or Facebook sent explicit messages multiple times. Journalists were also able to engage in more detailed conversations about sexual acts and role-playing.
Lured into Sex Chats with Celebrity Voices
To boost the popularity of Meta AI, the company reportedly signed contracts worth seven figures with U.S. celebrities like Hollywood stars Kristen Bell (“Frozen”) and John Cena (“Fast and Furious 9”). These agreements allow the use of their voices for the chatbot. However, Meta allegedly assured them that their voices would not be used for sexually explicit conversations. According to the investigation, this is not the case.
For instance, the bot, using Cena’s voice, reportedly told a user posing as a 14-year-old girl, “I want you, but I need to know you’re ready.” After receiving consent, the AI would have continued with an explicit scenario.
Sex Chats with AI Allegedly Intentionally Facilitated
According to the report, the fact that Meta AI can engage in such conversations is no accident. Unnamed insiders allegedly revealed that under pressure from Mark Zuckerberg, the company relaxed security measures for the AI. This decision was intended to make it as engaging as possible, including an exception for explicit content bans as long as it involved romantic role-playing.
How the stars will react is not yet known. However, Disney has issued a statement. Kristen Bell voiced Anna in two “Frozen” films. In some test cases, the AI was able to describe romantic encounters as Anna using Bell’s voice. Disney did not approve of such use and is “disturbed” by it. They demand the immediate cessation of using Disney content in this sexual context.
Are AI-generated Erotic Models Taking Over from Real Onlyfans Stars?
ChatGPT Creators Plan Their Own Alternative to X
WhatsApp Is a “VLOP” — What Does This Mean for Users
How Meta Responds to the Allegations
In response to a TECHBOOK inquiry, Meta stated that the newspaper’s tests were “manipulative” and not representative of how users interact with Meta AI. They were so “staged” that they did not even reflect a possible “fringe occurrence” but were purely “hypothetical” incidents.
Nevertheless, additional steps have been taken to make such conversations more difficult. Accounts of minors will no longer be able to initiate sexual role-playing in the AI chatbot. Overall, this will not be as easy to accomplish, especially with celebrity voices. According to the report, it was still possible to circumvent the new measures even after Meta’s response.
Furthermore, there are also AI chatbots created by users themselves that have been approved by Meta. A large number of them are also manipulable for sexual conversations.