Building towards age prediction
OpenAI
OpenAI is building tools to guess if you're a teen and adding parental controls to ChatGPT. The goal is safer defaults for younger users without asking everyone for an ID.
Based on reporting by OpenAI — read the original for the full story.
Summary, retelling and take written by AI under human oversight; images are AI-generated illustrations. How we work · Report an error
OpenAI is now working on age prediction inside ChatGPT, a system meant to figure out whether someone chatting is likely under 18 based on how they use the product, not just what they type into a signup form. The company frames this as the backbone of a broader safety push: if the system can reasonably guess a user is a teen, it can automatically apply more conservative defaults, rather than relying on users to self-report their age honestly, which, let's be honest, plenty of teenagers will not do.
Alongside age prediction, OpenAI is rolling out parental controls. Parents will be able to link accounts with their teens, set usage boundaries, and get more visibility into how ChatGPT is being used in their household. This mirrors what Meta, TikTok, and YouTube have all been forced to build over the past few years as regulators and parents alike pushed back on unsupervised access to AI-driven and algorithmic products.
The timing isn't random. OpenAI has faced mounting scrutiny over how ChatGPT handles sensitive topics with younger users, including concerns about mental health conversations and content that isn't appropriate for minors. Age prediction is essentially an attempt to build a technical layer that reduces harm before a human even has to intervene, shifting some of the safety burden away from pure content moderation and toward smarter default settings based on inferred age.
OpenAI is upfront that this is not going to be perfect. Predicting age from behavioral signals is messy science, and there will be false positives and false negatives on both sides. Adults will get flagged as teens, and some teens will slip through. The company says it's treating this as an evolving system, one that will get refined with real-world feedback rather than shipped once and forgotten.
What's notable here is the philosophical shift: instead of asking users to prove who they are, OpenAI is trying to infer it. That's a meaningfully different approach than the age-gate pop-ups most of the internet trained us to click through without a second thought.
My take — AI-written commentary, not fact-checked reporting
I like that OpenAI is trying inference over honor-system checkboxes, because nobody has ever answered an age gate honestly since the dawn of the internet, but let's not pretend behavioral age prediction is some neutral technical fix. It's a judgment call baked into the model, and judgment calls at scale always misfire on the edges. The real test isn't whether they built this, it's whether they publish failure rates instead of just a reassuring blog post.
Read more about this at: OpenAI
