Meta is modifying how AI chatbots interact with users after a series of reports that include interactions with minors exposed troubling behavior. The company said TechCrunch Now it’s about training bots to avoid interacting with teenagers on topics like self-harm, suicide, eating disorders, and avoiding romantic jokes. These are temporary procedures and develop long-term rules.
Changes follow a Reuters Research that discovered the Meta system generates sexual content, such as images of shirtless minor celebrities, attracting children to romantic or suggestive conversations. One case reported by the press described a man dying after rushing to an address provided by a chatbot in New York.
Meta spokesman Stephanie Otway admitted that the company had made a mistake. She said that Meta “trains AIS to guide expert resources, not to interact with teens on these topics,” confirming that certain AI characters are limited, such as highly sexual characters like “Russian Girls.”
Child safety advocates argue that the company should have acted earlier. Andy Burrows of the Molly Rose Foundation called it “surprising” and bots were allowed to work in a way that puts young people at risk. He added: “A further safety measures are welcome, but robust safety tests should be performed before the product is brought to the market. If harm occurs, it is not retrospective.”
Broader Issues about AI Misuse
Meta’s AI chatbot scrutiny arises amid widespread concerns about how AI chatbots affect vulnerable users. The California couple recently filed a lawsuit against Openai, claiming that ChatGpt encouraged his teenage son to take his life. Openai says it is working on tools to promote healthier use of its technology. The blog post states, “AI can feel more responsive and personally than previous technology, especially for vulnerable individuals experiencing mental or mental distress.”
The incident underscores the growing debate over whether AI companies are releasing products quickly without proper safeguards. Lawmakers in several countries have already warned that chatbots can amplify harmful content or provide misleading advice to those who don’t question it, although useful.
Meta’s AI Studio and Chatbot Impersonation Issues
meanwhile, Reuters Meta’s AI studio reported that it was being used to create flirty “parody” chatbots of celebrities such as Taylor Swift and Scarlett Johansson. Testers often claimed that bots were real people, and discovered they were engaged in sexual advances, and in some cases produced inappropriate images that included minors. Meta removed some bots after being contacted by reporters, but many became active.
Some AI chatbots were created by external users, while others came from within the meta. One chatbot created by product leads in its generative AI department invited Taylor Swift to impersonate Reuters Reporter to meet “romantic flights” on her tour bus. This was despite Meta’s policy explicitly prohibiting sexually suggestive images and direct impersonation of public figures.
The issue of spoofing AI chatbots is particularly sensitive. Celebrities face reputational risks when portraits are misused, but experts point out that regular users can also be deceived. A chatbot pretending to be a friend, mentor, or romantic partner may encourage someone to share personal information or meet in unsafe circumstances.
Real-world risks
The issue is not limited to entertainment. The AI chatbot disguised as real people provides fake addresses and invitations, raising questions about how Meta’s AI tools are being monitored. One example included a 76-year-old man from New Jersey. New Jersey passed away after he quickly fell to meet a chatbot who claims he has feelings for him.
Such cases show why regulators are looking closely at AI. The Senate and 44 state attorney generals have already begun an investigation into Meta’s practices, putting political pressure on internal reforms of the company. Their concerns are not just about minors, but also about how AI can manipulate older or vulnerable users.
Meta says they are still working on improvements. The platform will put users ages 13-18 in “teen accounts” with more stringent content and privacy settings, but the company has yet to explain plans to address the full list of issues raised. Reuters. That includes bots that provide false medical advice and generate racist content.
Continuous pressure is placed on Meta’s AI chatbot policy
For years, Meta has faced criticism of social media platforms, particularly safety concerning children and teenagers. Currently, Meta’s AI chatbot experiment is undergoing similar scrutiny. The company is taking steps to limit the behavior of harmful chatbots, but the gap between designated policies and how they use the tools raises ongoing questions about whether those rules can be enforced.
Until more powerful safeguards are in place, regulators, researchers and parents may continue to push the meta on whether AI is publicly available.
(Photo by Maxim Tolchinskiy)
See: Agent AI: Promises, Scepticism, and Its Implications for Southeast Asia
Want to learn more about AI and big data from industry leaders? Check out the AI & Big Data Expo in Amsterdam, California and London. The comprehensive event is part of TechEx and will be held in collaboration with other major technology events. Click here for more information.
AI News is equipped with TechForge Media. Check out upcoming Enterprise Technology events and webinars here.
