Educator, author and trusted guide to families navigating AI
Dan Fitzpatrick writes about AI in education for Forbes and is the author of The Educators' AI Guide 2026, The AI Classroom and Infinite Education. He founded aieducator.tools after working as a teacher, school leader and education strategist.
He also hosts the AI for Educators Daily podcast and advises schools and education bodies in the UK and abroad.
Your teen's own ChatGPT or Google account is not something the school can open. But a school laptop and the school Wi-Fi change the answer. Here is the account, device and network rule, a table you can keep, and a two-minute check to do together tonight.
If your child's laptop and their account both belong to the school, Google says plainly that parents cannot add controls. Here is how to tell which of three setups you have, what you can still control, and what to ask the school about AI on a device it manages.
Four AI settings you can genuinely switch off tonight, on iPhone, Android, Snapchat and Instagram, and the honest list of what has no off switch at all, starting with the AI answers at the top of every Google search.
Nobody can tell you which jobs will exist, and the OECD has stopped trying. What its 2026 exposure work can name is the kind of work machines are still worst at, and those four capacities are the ones a parent can build at home, age by age, long before anyone knows the job.
Replika's terms of service put the app at 18 and over, and Apple rates it 18+, so for a teenager this is not a settings decision. What two named assessments found, why a European regulator called the age check deficient, and the exact steps to block it on an iPhone or an Android phone tonight.
A straight answer, age by age, from someone who has sat in the safeguarding meetings. What ChatGPT now does for 13 to 17 year olds, what parental controls will and will not reach, and the two minutes tonight that matter more than any setting.
A calm read of September 2026 for families: thirteen new California child safety laws, Europe's KIDS Act proposal, a homework tool rated unacceptable risk for teenagers, and the dates that decide when any of it reaches your home.
Before eight, the AI a child meets is usually the voice assistant already running in the house, not a chatbot they chose. Here is what to allow, what to hold back, and the one thing to change on the device at each stage from toddler to seven.
There is no age that makes a child ready for AI, and the official guidance is a floor rather than an answer. What Common Sense Media, UNESCO and the AI companies actually say, why most children start before a parent decides, and the three signs that tell you whether yours is ready.
ChatGPT keeps your child's chats, uploads, date of birth, device and approximate location, and on a personal account those chats can help train the model. Two settings change that, and a school-managed account reverses the whole trade.
Roblox now age-checks every player before they can chat, which is a real improvement. But the AI characters inside games are built by developers, and Roblox's own parental controls page says your chat settings do not reach them. Here is what does cover them, the one question to ask your child, and what to do at each age.
A former Google trust and safety chief says AI could harm children more than social media did. Here is what the evidence actually supports, why the two are different problems, and what that changes for you at home.
