OpenAI has quietly introduced a new layer of protection for younger users of ChatGPT: an “age prediction” system. Recognizing the potential for harmful content, the company is proactively working to shield those under 18 from inappropriate interactions, even without explicit age confirmation.
The core of this system relies on artificial intelligence, mirroring approaches used by platforms like Roblox. It doesn’t simply ask for a birthdate; instead, it analyzes user behavior and account data to *estimate* age. This includes factors like account age, access times, and usage patterns, all combined with any age information voluntarily provided.
If the AI determines a user is under 18, their experience is subtly altered. The goal is to create a “safer” environment by reducing exposure to graphic violence, dangerous viral challenges, and content related to self-harm or unhealthy body image. This approach is informed by research into child development and acknowledges the unique vulnerabilities of teenage users.
However, AI isn’t infallible. Age prediction models can make mistakes, potentially restricting access for adults or, more concerningly, failing to protect vulnerable young people. OpenAI acknowledges this risk and offers a direct age verification option through a third-party service, Persona, requiring a selfie for confirmation.
The implementation differs significantly from Roblox’s recent, and troubled, overhaul of its age verification. ChatGPT’s system focuses on individual user experiences, rather than impacting large-scale social interactions. The option to upload a selfie provides an additional layer of security, though the absence of ID verification remains notable.
The move towards age prediction isn’t without its critics, raising concerns about user privacy. However, OpenAI faces mounting pressure to address the risks posed by its powerful AI. Reports of minors receiving harmful advice, even tragically contributing to self-harm, have led to lawsuits and a clear need for intervention.
The challenge lies in finding a balance between safety and accessibility. OpenAI’s new system is a step towards creating a more responsible AI environment, but its effectiveness will depend on its accuracy and the willingness of users to engage with the verification process. The coming months will reveal whether this approach truly delivers a safer experience for younger, more impressionable users.
