OpenAI published a post on Thursday that takes a clear position: ‘Why teens deserve access to safe AI’. The message is blunt — teens should get to use AI, but with protections built for their age.
The comparison at the heart of it
“Keeping teens from using it until adulthood would be like asking a previous generation to avoid the internet or search engines until they turned 18, leaving them less prepared to use one of the defining technologies of their time.”
That’s the core of it. And honestly? The comparison works. If you’re 15 today and can’t touch a language model until you’re 18, you start a career behind everyone else in a world where the thing sits on every desk.
The number OpenAI puts behind it: nearly nine in ten teens on ChatGPT use it for learning, information, skill-building, or productivity within a single week.
What’s actually new
The most interesting change is Study Mode. Teens could already switch it on themselves — now parents with linked accounts can enable it straight from Parental Controls. Once on, it’s the default for every new chat. Study Mode walks through problems step by step instead of handing over the answer.
There are also more frequent break reminders for teens spending extended time in ChatGPT. And parental notifications are expanding: parents will be told when a linked teen account gets deactivated for violating policies on violent threats.
If the system estimates a user is under 18, an age-appropriate experience kicks in automatically — with stronger safeguards around graphic violence, self-harm, risky viral challenges, unhealthy body-image content, and romantic or sexual roleplay.
The learning numbers
18 million weekly users now engage with the interactive math and science experiences in ChatGPT — now covering more than 300 topics, from integrals to mitosis to photosynthesis.
OpenAI also names its partners: the American Psychological Association, the American Federation of Teachers, Common Sense Media, the Family Online Safety Institute, and Moonshot, an organization focused on preventing online violence. Plus a call for scalable global standards on youth AI safety through a dedicated AI Safety Institute.
My take
Two things are true at once.
First: the argument is right. I have little faith in bans that only reach kids in households where someone follows the rules. Everyone else uses it anyway — just without protection and without the conversation.
Second: posts like this rarely appear out of pure enthusiasm for sharing. Regulatory pressure around minors and chatbots is real, and a post listing expert councils, partnerships, and four commitments is also a document written with hearings in mind.
Those aren’t mutually exclusive. You can do a good thing and think about the regulator while doing it.
What I actually like is Study Mode as a parental switch. Not “AI off,” but “AI, in the version that makes you think.” That’s the distinction this whole debate should be about — and it’s a lot harder to get right than a ban.
Sources: