Children can have less power, less context and fewer practical ways to challenge an automated decision. Q10 uses a child-first test to ask whether a system remains understandable, reversible and non-exploitative for people who are especially vulnerable.
What information about a child is collected? Why is it needed? Who can see it? Can the child or guardian understand what the AI is doing? Is there a real human appeal path? Could the system manipulate, shame, exclude or create dependency? Can a harmful action be stopped or reversed?
It is also a design stress test. If a system only works safely for powerful expert users, that is important information about the system.
Open the Child First AI Checklist · Browse people, family and health resources.