OpenAI has unveiled a significant new safety feature for its popular AI chatbot, ChatGPT, designed to protect younger users from inappropriate material. The age prediction tool, which launched globally this week, analyses user behaviour to estimate whether an individual is under the age of 18, automatically applying content restrictions if a minor is detected.
How the Age Prediction System Operates
The innovative feature works by continuously monitoring a person's usage patterns within the ChatGPT platform. These behavioural signals include detailed metrics such as how and when an account is actively used, alongside the duration the account has been in existence. By analysing this data, OpenAI's sophisticated prediction model can make an informed estimation regarding the user's age bracket.
Automatic Safeguards for Younger Users
If the system determines that a user is likely a minor, ChatGPT will immediately implement a series of protective measures designed to limit exposure to sensitive or mature content. This safeguarded content encompasses a broad range of material, including sexually explicit or violent depictions, representations of self-harm, content promoting unhealthy lifestyles, and viral challenges that encourage risky or harmful behaviours.
"This approach is guided by expert input and rooted in academic literature about the science of child development," OpenAI explained in an official website post. "It acknowledges known teen differences in risk perception, impulse control, peer influence, and emotional regulation." The company further emphasised its commitment to continually refining these protections to counteract any attempts to bypass the established safeguards.
Verification Options for Incorrect Classifications
Recognising that the automated system may not be flawless, OpenAI has established a clear verification pathway for users who believe they have been incorrectly identified as underage. Affected individuals will have two primary options to confirm their actual age. They can upload a government-issued identification document, such as a driver's licence or passport. Alternatively, they may submit a selfie through a third-party identity verification service called Persona, which utilises artificial intelligence to assess and confirm a person's age.
Rollout and Regional Considerations
The new age prediction functionality is being deployed to users worldwide this week. However, OpenAI has noted that there may be a slight implementation delay for users within the European Union, due to specific regional regulatory requirements and compliance checks that must be addressed.
Anticipated Introduction of "Adult Mode"
This safety feature arrives as a precursor to a highly anticipated "adult mode" for ChatGPT, which is scheduled for release in the first quarter of 2026, according to statements made by OpenAI's app chief, Fidji Simo, in December. The forthcoming mode is expected to grant users the ability to generate adult-oriented content, including erotica and violent depictions, using ChatGPT's text, image, and video generation tools.
It will also permit discussions on sensitive topics that are currently restricted on the platform. Importantly, OpenAI has clarified that even within adult mode, strict prohibitions will remain in place against non-consensual content, deepfakes, and material depicting or encouraging illegal activities, maintaining a baseline of ethical use.
