Meta AI introduces parent alerts for teen self-harm risks

Meta announced on Friday that parents utilising its family supervision tools will now receive alerts if their teenager's conversations with Meta AI suggest they may be at risk of suicide or self-harm. This safety feature applies to chats on Instagram, Facebook, and Meta Horizon, expanding upon a monitoring system launched earlier this year. The previous iteration flagged search queries for concerning terms, whereas the updated software actively analyses interactive conversational logs.

The social media giant published the update on its official website, detailing how the notification process operates. When the assistant detects thoughts of self-harm, it automatically provides hotline contacts and advises the teenager to speak with a trusted adult. Under the new protocol, the system simultaneously transmits a notification to registered parents or supervisors if the signals indicate genuine danger. To ensure accuracy, human moderators manually review every flagged interaction before dispatching the final alert.

The company acknowledged that the double-checking process might occasionally trigger notifications when no real danger exists, but stated this conservative threshold is the correct starting point. The protective system is strictly limited to accounts enrolled in the family supervision programme through the Meta Family Centre. Currently, the utility is active only in the United Kingdom, the United States, Canada, and Australia, with no immediate international expansion plans confirmed.

To protect teenagers' privacy, the outgoing alerts do not disclose the transcript or specific wording of the conversation. Instead, the system prompts parents to initiate a supportive dialogue and attaches expert guidelines explaining how to approach the sensitive topic. Additionally, the technology firm is developing a mechanism to alert emergency services directly if an adult or teenager indicates an imminent threat to life. This expansion mirrors an existing framework for standard Facebook and Instagram posts, which generated over 19,000 emergency responder wellness checks globally last year.