Filed under
Practical guidance on privacy, online safety, deepfakes, AI companions and digital wellbeing.
This section collects 21 pieces on Safety and Wellbeing, including reviews, reported notes and practical guidance.
21 pieces
Your child, or your child's school friend, wants to know whether their school can see what they type into ChatGPT. Usually not. A personal account on a personal phone is private to them. A ChatGPT account the school owns is not, and a school laptop reads what is typed on any site. The school email address is not what decides it.
My AI sits above your child's friends in Snapchat and nobody chose to install it. Here is what Snapchat's own help pages say it keeps, how to delete it, what two named assessments found, and the difference between hiding My AI and stopping it.
Parents type this at 10pm and get either a shrug or a scare. Here is the honest answer: there is no diagnosis called AI addiction and no published hours-a-day limit, so the useful question is what the time replaced, not how much of it there was.
The AI age limits are all different, and not one of them is a lock. What ChatGPT, Gemini, Claude and Character.AI actually require, why the number is usually 13, and the two-minute check that tells you which version of the app your child is getting.
Chatbots built for children exist, and for under-13s they are the right place to start. But a filter is a fence around one app, not around your child. What the safe ones fix, what no setting can guarantee, and the question worth asking tonight.
Three independent tests of AI toys, a four-year California moratorium now signed into law but not starting until 2027, and the question none of them asks: what a toy that answers everything does to the conversations your child would otherwise be having with you.
In August 2026 Google switched Gemini on by default inside Google Classroom for students of every age, and Common Sense Media rates it High Risk. Which Gemini your child has, what the rating actually means, four questions to ask the school, and what to do at home by age.
ChatGPT's parental controls are worth setting and cover about a third of the problem. Checked October 2026: what starts on and off, what no parent can see, and the two layers the app cannot reach.
Character.AI ended open-ended chat for under-18s in November 2025, so the real question is whether your child talks to any AI as a friend. A former assistant head on when to worry and what to say.
