Did you know there is a ChatGPT for teens? No? Don’t worry, neither did the teens.
OpenAI rolled it out this week, and it’s a separate version of the chatbot specifically for 13- to 17-year-olds. What was interesting to me was that it was a pretty quiet launch.
A press release, a run of headlines on the tech sites, and then nothing. No campaign aimed at parents, no email to the accounts with kids sitting on them.
My two teenagers didn’t even know it happened. I doubt you’ll get a different answer in your house.
So here’s my take, plainly. OpenAI pushed this out because they had to, not because they wanted to. Quiet rollout, box ticked, on to the next thing. This to me feels more like a legal thing trying to fend off lawsuits than it is for the good of young people.
Still interested in what it even is? It’s basically ChatGPT for 13 to 17 year olds and you don’t opt in, so if your kid uses ChatGPT they’re probably already in it. As a parent, you can link your account to theirs, shut it off at set hours, and make it answer homework with questions instead of handing over the essay.
You can’t see a word they say to it. No transcript, no summary, no heads up if a chat goes somewhere it shouldn’t.
That’s the basic story. Two switches and a lot of trust.
The reason to care isn’t the features. It’s that your kid got moved into a different product and nobody thought to mention it to you.
And it’s already going sideways. The age detection is misfiring badly enough that paying adult customers are stuck in teen mode, and getting out means verifying yourself through a third party firm that can ask for government ID or a selfie video. Meanwhile the kid who registered at twelve claiming to be twenty-two was never caught at all.
The internet has been asking kids to confirm they’re 13 since about 1998. I’m not sure it’s ever caught on. For anyone who has had to deal with Roblox, or Nintendo Switch accounts, you know the age thing can cause a lot of friction.
I’m not the only one unconvinced about this new rollout of ChatGPT. Fairplay, a children’s advocacy group, called it a step in the right direction and then warned about “a false sense of safety based on the announcement”. They also flagged that some of the protections that actually matter, like limiting what it remembers about your kid, sit switched off unless you go hunting for them.
So do these companies care about teenagers? ChatGPT got to something like 900 million weekly users, and a stack of lawsuits, before anybody built a version for the kids who were already in there.
If they cared, the protections that matter would be on by default.
And parents would have heard about this from OpenAI, instead of from a blog.
So no, don’t ban it and don’t panic. Go and look at the settings, they’re better than nothing. Just know that a launch this quiet is what it looks like when a company does the minimum and hopes nobody checks.
I guess the positive is there is a company starting the conversation at least. I wonder if the other AI tech giants will follow suit. ChatGPT, Claude and Gemini are going to be a huge part of a teen’s life going forward and anything that moves in the right direction is good. I just hope our many levels of Government will catch up and do what they should to protect us all.
For more opinions and thoughts on navigating the wonderful world of teens, check out my parent hub.
Honestly, the “did you even know it existed” part is the whole story. We found out about the teen account the same way we find out about most things in this house: our 15-year-old mentioned it sideways while we were scraping the grate after a pork butt cook. No email, no banner, nothing in the school newsletter. Just a kid who already had it set up and a dad standing there with a wire brush.
What surprised me is how little the two parental settings actually cover. They handle the obvious stuff, but the day-to-day questions are the ones nobody writes a toggle for: is it okay to use it to outline an essay, is it cheating to have it explain a math step, what happens when it confidently gets something wrong and he repeats it at dinner like it’s fact. We ended up doing what you described – sitting down and talking instead of locking it down, because the lockdown version just moves the activity to a friend’s phone.
One thing that did help us was making the AI stuff visible instead of secret. My son is into design, so we started messing around together with an ai image generator to make ridiculous fake album covers for our BBQ playlists – the one we use is https://soralum.com/image-ai/image-generator – and it turned into a weirdly good teaching moment. He could see immediately that the output is only as good as the prompt, that it invents details, and that “it looks real” and “it is real” are not the same sentence. Way easier to explain hallucination when the thing has just given you a brisket with six ends.
The other bit of advice I’d add for other parents: check the account age settings yourself rather than trusting the marketing page, and keep the conversation about *why* rather than *how much*. Screen time limits never worked here. Asking him to show me what he made, and asking him what he thinks it got wrong, works every single time.
Anyway, thanks for writing this one up. Also, now I want to know what you’re cooking this weekend.