A parental risk alert is an escalation feature for high-severity situations such as indications of acute self-harm or violence risk. Detection can come from automated analysis, but the alert threshold, recipient, content, timing, and follow-up process require careful design.
The feature creates a difficult balance among safety, privacy, teenage autonomy, false alarms, missed warnings, and legal duties. Users should understand when alerts may occur, and the system should disclose only information necessary for the defined safety response.
ELI5
A parental risk alert is a warning sent to a connected parent or guardian when a young person's AI conversation suggests an immediate and serious danger. It is meant for urgent safety situations, not ordinary private questions.
For example, repeated messages showing an acute risk of self-harm could trigger an alert that encourages a trusted adult to check on the teenager. The system needs careful thresholds because a false alarm can harm trust, while a missed warning can leave someone without help.
