OpenAI's ChatGPT Adds Age Prediction for Youth Safety

OpenAI is enhancing its ChatGPT platform with a new age prediction model aimed at safeguarding younger users. This initiative, announced Tuesday, is designed to identify accounts likely belonging to individuals under 18, allowing for the implementation of protective measures.

How the Age Prediction Works

The sophisticated model analyzes a combination of account-level and behavioral data. This includes examining usage patterns over time, the duration an account has been active, the typical times a user engages with the platform, and the age information self-reported by the user.

Addressing Safety Concerns and Regulatory Scrutiny

This development comes as OpenAI faces increased attention regarding user safety, particularly concerning minors. The AI startup, alongside other technology firms, is currently under investigation by the FTC concerning the potential impact of AI chatbots on children and teenagers. OpenAI also faces several lawsuits, including one linked to the tragic suicide of a teenage boy.

Implementing Protections for Underage Users

When the age prediction model flags an account as potentially belonging to a user under 18, ChatGPT will automatically activate features to limit exposure to sensitive content, such as depictions of self-harm. OpenAI emphasizes that these measures are intended to create a safer online environment.

Restoring Access for Misidentified Users

Recognizing that the model may not always be perfectly accurate, OpenAI has established a recourse for users who are incorrectly identified as underage. These individuals can utilize the identity-verification service, Persona, to regain full access to the platform. Persona is a tool also employed by companies like Roblox, which has similarly been urged by lawmakers to bolster its child protection protocols.

A Continued Commitment to Safety

OpenAI's rollout of the age prediction model is part of a broader strategy to enhance user safety. In August, the company announced its intention to introduce parental controls, empowering parents to better understand and manage their teenagers' engagement with ChatGPT. These parental controls were subsequently launched in September, coinciding with the announcement that an age prediction system was in development.

Furthermore, OpenAI has proactively engaged with experts, forming a council of eight professionals in October. This council is tasked with providing crucial insights into the potential effects of AI on users' mental well-being, emotional states, and motivation.

OpenAI has stated its ongoing commitment to refining the accuracy of its age prediction model. The system is slated for deployment in the European Union within the coming weeks, ensuring compliance with regional regulations.

For immediate support if you are experiencing suicidal thoughts or are in distress, please contact the Suicide & Crisis Lifeline at 988. Trained counselors are available to provide assistance.