ChatGPT Implements Age Prediction to Enhance Teen Safety
OpenAI has introduced new age prediction capabilities to its flagship chatbot, ChatGPT, in a significant move aimed at bolstering safety for teenage users. This initiative replaces the previous method of simply ticking a box to confirm being over 18, with a more sophisticated system that analyses behavioural signals to estimate a user's age.
How Age Prediction Works in ChatGPT
The age prediction feature does not activate immediately upon account creation. Instead, it operates in the background, monitoring various factors such as the duration of account ownership and activity patterns. For instance, if usage aligns with typical school hours, it might indicate a younger user. OpenAI emphasises that this approach allows for a more nuanced assessment of age without intrusive initial checks.
If ChatGPT determines that a user is under 18, it will automatically restrict access to certain types of content deemed harmful. This includes material related to gore, risky viral challenges, eating disorders, and adult content. The goal is to create a safer digital environment for minors while maintaining functionality for adults.
Verification Options for Users
For users who are incorrectly identified as underage, OpenAI provides a verification process. This involves uploading a scan of a government-issued ID or a selfie, which is then analysed by a third-party identity-verification service. In a recent test by a Metro reporter, the system requested access to a laptop camera to film the user, asking them to look directly at the camera and then turn left and right for age estimation.
Upon successful verification, users receive a confirmation message stating: 'You'll soon be able to use all ChatGPT features available to adults over 18. It may take a bit for everything to refresh.' Additionally, if a user self-reports their age as under 18, safeguards are immediately applied to prevent exposure to inappropriate content.
Context and Industry Trends
This development follows OpenAI's introduction of parental controls last year, which allow guardians to oversee their teens' accounts and receive notifications about potential signs of distress. The age prediction feature has been requested by OpenAI's developer community for over a year, highlighting the growing need for tools to support digital parenting in an era of widespread AI usage.
ChatGPT is not alone in adopting age-checking technologies, often referred to as age-gating. Major platforms like Instagram, YouTube, TikTok, and Roblox have also implemented restrictions for children's accounts in recent months. For example, Instagram uses AI to analyse messages for birthday references and interactions with other underage users to estimate age.
Regional Rollout and Future Plans
The age prediction feature is currently being rolled out globally, with a slight delay in the European Union to accommodate regional regulatory requirements. OpenAI plans to introduce it there within a few weeks. In a blog post, the company stated: 'While this is an important milestone, our work to support teen safety is ongoing.' They are collaborating with experts, including the American Psychological Association and ConnectSafely, to refine these measures.
This move comes amid broader concerns about the impact of AI on young people. Recent studies indicate that more than half of UK teens have used AI for emotional support, raising alarms among mental health professionals who caution that chatbots like ChatGPT are not substitutes for therapy. Past incidents, such as a glitch allowing underage users to access erotic material, have underscored the need for robust safety features.
As governments consider stricter regulations, such as Australia's ban on social media accounts for youngsters, OpenAI's proactive steps reflect a commitment to responsible AI development. The integration of age prediction and verification aims to balance user safety with the freedom for adults to utilise ChatGPT's capabilities within safe boundaries.
