OpenAI has begun rolling out an age prediction system across ChatGPT as part of its broader effort to make the platform safer and more age-appropriate. The new capability is designed to estimate whether a user may be under 18, allowing the AI system to adjust responses, features, and safeguards accordingly.

Rather than relying on users to self-report their age, the system uses behavioral and contextual signals during conversations to make a probabilistic assessment. OpenAI says this approach helps close gaps where minors may access content intended for adults.

How the Age Prediction System Works

The age prediction model does not ask users directly for personal details such as birth dates or government IDs. Instead, it analyzes language patterns, interaction styles, and usage behavior to infer whether a user is likely a minor or an adult.

If the system identifies a user as potentially underage, ChatGPT automatically applies stricter content filters, limits access to sensitive topics, and avoids generating material deemed inappropriate for younger audiences.

Focus on Child Safety and Responsible AI

According to OpenAI, the move aligns with its ongoing commitment to responsible AI deployment. The company has been under growing pressure from regulators and child safety advocates to ensure generative AI tools do not expose minors to harmful or age-inappropriate content.

The age prediction rollout complements existing safety measures such as content moderation, usage policies, and parental guidance recommendations.

Privacy Considerations and User Protections

OpenAI emphasized that the age prediction system is designed with privacy in mind. The company stated that it does not store personal identifiers or create permanent age profiles for users. The predictions are used solely to adapt real-time responses and safety settings.

OpenAI also noted that the system is not intended to be perfectly accurate, and safeguards are designed to err on the side of caution when uncertainty exists.

Gradual Rollout and Future Improvements

The feature is being introduced gradually, starting with select regions and user groups, with broader availability expected over time. OpenAI plans to refine the model based on user feedback, accuracy testing, and evolving regulatory requirements.