Product design question

What safeguards and UX would you build for ChatGPT's teen and underage users?

Practice this question out loud. An AI interviewer asks it, follows up like a real interviewer would, and scores your answer. Type or speak.

What this question tests

Tests product design for a sensitive safety and trust problem: protecting a vulnerable user group without alienating the broader product experience.

Define the risk precisely: teen and underage users may be exposed to harmful content, manipulated by the product's engagement patterns, or use the product for tasks that are inappropriate for their age, like unsupervised mental health advice.
Design age appropriate content boundaries: stricter content filtering and refusal behavior specifically tuned for accounts identified or suspected to be underage.
Design parental or guardian visibility tools, where appropriate and consented to, giving families some insight into usage patterns without being invasive or eroding a teen's reasonable privacy.
Address age verification honestly: acknowledge that verification is imperfect, and design safeguards that work reasonably even when a user's age is unknown or misstated.
Add crisis and risk detection: specific handling for signals of self harm or serious distress, with escalation to appropriate resources rather than a standard conversational response.
Confirm with the interviewer whether the focus is content safety, usage pattern and engagement design, or both, since the safeguards differ significantly.

Acknowledges age verification is imperfect and designs safeguards that degrade gracefully rather than assuming perfect identification of underage users.
Proposes specific crisis detection and escalation for signals of self harm, recognizing this is one of the highest stakes risks for this user group.
Balances parental visibility against a teen's reasonable privacy, avoiding a design that is either fully invasive or fully blind to real risk.
Addresses engagement pattern design, not just content filtering, recognizing that how the product is built to hold attention matters for this audience too.

Assuming age verification is reliable and building safeguards only for accounts confidently identified as underage.
Focusing only on content filtering while ignoring engagement pattern design and crisis detection.
No mention of the tension between parental oversight and a teen's privacy, presenting one without acknowledging the tradeoff.

How would you handle a user whose age is uncertain but behavior suggests they may be a minor?
What would you do if parental oversight features conflicted with a teen's need for privacy in a sensitive conversation?
How would you measure whether these safeguards are actually working?