Meta is expanding parental alerts about possible teen suicide or self-harm risk to conversations with Meta AI on Facebook and Messenger. Its 29 October update also adds alerts for repeated searches on Facebook, extending features previously announced for Instagram.

Meta says the alerts are rolling out globally to supervising parents. That qualification matters: the announcement does not promise automatic coverage for every parent and teenager, or establish that the rollout is complete for every eligible account.

AI conversations and repeated searches have different triggers

The October update describes two routes to an alert:

  • A teen repeatedly searches for terms related to suicide or self-harm on Facebook or Instagram within a short period.
  • A conversation with Meta AI on Instagram, Facebook or Messenger suggests the teen may be at risk of suicide or self-harm.

The second route concerns chats with Meta’s AI assistant. It does not describe a system scanning every private conversation between people.

For search alerts, Meta does not specify the number of searches, the length of the time window or a complete list of triggering terms. The announcement also does not say that these search alerts undergo human review.

Meta says flagged AI chats receive manual review

The updated page retains the AI-chat process described in Meta’s original 16 July announcement. Meta says a dedicated AI system identifies concerning conversations, and that all AI-flagged chats are manually reviewed before an alert is sent to a parent.

Human review does not remove ambiguity. Meta acknowledges that it may send an alert when intent is unclear, including situations where there may be no real cause for concern. It gives no measured rate for those alerts.

What parents receive, and what remains unclear

Meta says parents receive an alert and expert resources. Its announcement does not establish whether the notification includes excerpts or whether parents can access a conversation transcript.

The announcement does not spell out reviewers’ access to wider conversation history, their detailed permissions or how long that material is retained.

The October update identifies supervising parents as recipients, but does not provide a platform-by-platform account of age eligibility, consent or setup requirements. We could not retrieve substantive Help Centre guidance to verify those details. Parents should not assume that an Instagram supervision arrangement establishes the same requirements or coverage on Facebook and Messenger.

Detection results and emergency-service plans remain unverified

Meta’s announcement describes how the system is intended to work. It supplies no detection-accuracy figures, missed-alert or false-alert rates, or evaluation of outcomes. No independent detection or outcome evidence was retrieved for this article, so the available material does not establish how reliably the alerts identify risk or help teenagers.

Emergency-service contact is a separate feature. The updated announcement still describes Meta as building the ability for Meta AI to alert emergency services when an adult or teen may face imminent suicide risk. The October update does not announce that capability’s launch; we could not verify a more recent status from other sources.