If a teenager in your house uses ChatGPT for homework, the version they open just changed. On 18 August, OpenAI began rolling out ChatGPT for Teens, a separate experience with tighter content limits, parental controls and a study mode that pushes users to work through a problem instead of copying the answer. It landed in the same month Malaysia started enforcing its own rules for minors online, and the two approaches could hardly be further apart.
Table of Contents
What OpenAI actually changed
According to TechCrunch, the teen version switches on automatically when an account says it belongs to someone aged 13 to 17, or when OpenAI's systems estimate the user is under 18 from how they type and what they ask. It restricts what the assistant will discuss, adds parental controls that let a parent link their account and adjust settings, and introduces a Study Mode that offers step-by-step guidance rather than finished essays. It is available on both free and paid personal plans.
The key word is estimate. OpenAI is not checking anyone's identity document. It predicts age from behaviour, and where it is unsure it says it defaults to the teen experience. That is a softer bar than it sounds. A teenager who enters an adult birthday, or whose usage simply reads as older, can still end up in the standard product.
Malaysia already took the harder route
Malaysia's rules point the other way. Since 1 June, under the Online Safety Act 2025, platforms with large local audiences must verify user ages against government records such as MyKad or a passport, and keep under-16s off social media entirely. The Star reported in August that the communications ministry treats this verification as mandatory under the Child Protection Code, with penalties reaching RM10 million. Where OpenAI predicts, Malaysia demands proof.
That gap raises an awkward question for parents here. The law was written for social media, and it covers the platforms by size: the Facebooks, Instagrams, TikToks and YouTubes with millions of Malaysian users. An AI chatbot is not obviously any of those. So the tool a growing number of Malaysian students now lean on for schoolwork may sit outside the very rules built to protect them, while the apps they scroll for fun face hard identity checks. The same young users are heavy on those platforms too: a recent survey found most Malaysians turn to TikTok to learn and connect. The mismatch is nobody's fault in particular, but it is real.
Why this matters beyond the homework
Malaysia has spent the past year trying to write rules for fast-moving technology, from social media age checks to a framework for AI itself. The teen-safety question shows how quickly those rules can be overtaken. Regulators drew a careful line around one category of app, and the thing children adopted next was a different category that answers back in full sentences. Civil society groups have also warned that forcing people to hand government ID to every platform carries its own privacy risk, which is part of why a prediction-based model, for all its holes, is not simply the weaker choice.
For a Malaysian family, the practical takeaway is smaller than the policy debate. ChatGPT for Teens is worth switching on if a teenager in the house uses it, and the parental controls are worth the five minutes to set up. Just do not mistake a safer default for a locked door. The system is guessing your child's age, not verifying it, and no local rule yet forces it to do more.
Images courtesy of Richard Williams and Soundtrap on Unsplash.
