Meta has introduced a new safety feature for underage users, strengthening its existing parental supervision system. Starting today, when a teenager interacts with Meta AI and the content of the chat suggests thoughts of self-harm or suicide, the platform proactively sends a notification to the teen's parents or guardians. This follows the rollout a few months ago of the ability for parents to view the topics of their children's AI conversations. Meta aims to demonstrate that its digital environments can be safe for young people, a message directed at both families and regulators.
A dedicated AI system identifies risky conversations
Meta has developed a specific artificial intelligence model to recognize speech indicating potential danger. Before a notification is sent, a human reviewer manually examines all flagged chats. The goal is to avoid false alarms that could cause unnecessary panic. However, the company stated that in case of ambiguity, caution will prevail: if the human reviewer detects a doubt, the notification will still be sent. The feature is already live in the United States, United Kingdom, Australia, and Canada for parents using Instagram's parental supervision. By the end of the year, it will be extended to all users who have activated such controls worldwide.
Sponsored Protocol
Meta also provides resources and advice to parents
In addition to the notification, Meta supplies parents with resources and suggestions on how to handle the situation. The platform already directs the minor to crisis helplines and encourages them to talk to an adult. This new measure adds another layer of support by directly involving guardians. Meta also announced the development of a feature that will, in the future, allow contacting emergency services if a conversation with Meta AI suggests an imminent risk of suicide, for both minors and adults. In May, OpenAI launched a similar feature called "Trusted Contact for ChatGPT," which allows users to designate a trusted contact to be alerted in case of danger.
Sponsored Protocol
Another update concerns the "Limited Content" setting on Instagram: if a parent has activated this option to filter sensitive content (such as mature images or references to self-harm), their child's interactions with Meta AI will also be more restricted. For more on teen safety on social platforms, read the article on Meta alerting parents when teens discuss suicide with its AI chatbot. Meanwhile, OpenAI built GPT-Red, an LLM super-hacker for safety testing. For broader context on content moderation challenges, see the article on router disposal in Australia.
For more information on Meta's guidelines, visit the official teen safety page.
Source: https://www.engadget.com/2216412/meta-ai-alert-parents-if-teens-discuss-self-harm
