ChatGPT enables AI age prediction, and underage users will receive automatic protection

OpenAI is strengthening the content safety mechanisms of ChatGPT. According to the latest news, ChatGPT has implemented an age prediction feature worldwide (which will be launched in the EU region in the coming weeks) to automatically identify whether an account may belong to a user under 18, thereby enabling corresponding content filtering and usage restrictions. This is a practical application of AI technology in platform governance and also reflects the proactive response of generative AI products under regulatory pressure.

How the Age Prediction System Works

Recognition Principles

The system estimates user age by analyzing multiple behavioral signals, including account activity duration, usage patterns, interaction features, and more. This behavior-based inference method avoids the cumbersome process of mandatory identity verification while maintaining a certain level of privacy.

Protection Mechanisms

Once a user is identified as a minor, the system automatically activates two layers of protection:

• Age-appropriate content filtering to restrict access to content that may not be suitable for minors
• Usage restrictions, which may include adjustments to feature access permissions

Handling Misidentification

Adult users mistakenly identified as minors can restore full permissions through an identity verification process in the settings. This design balances protection and convenience.

Why This Is Important

Regulatory Context

Generative AI products face regulatory scrutiny worldwide, with child protection being a core concern. The EU’s Digital Services Act and various US state child privacy laws explicitly require platforms to fulfill protective obligations. OpenAI’s initiative is a proactive response to these regulatory pressures.

Industry Significance

AI-driven age recognition systems mark a shift from passive review to proactive prevention in platform governance. Compared to traditional manual review or rule-based filtering, this approach is more scalable. If effective, it is likely to become an industry standard, encouraging other generative AI platforms to follow suit.

Technological Innovation

Using behavioral analysis instead of identity verification to infer age is also reflected in the collaboration between Tinder and Worldcoin. Both aim to balance user privacy protection and security, but they take different technical paths—one relies on behavioral data, the other on biometric recognition.

Issues to Watch

Accuracy and Fairness

The accuracy of behavioral analysis directly impacts user experience. A high false-positive rate could require many adult users to repeatedly verify their identity, while a high false-negative rate might allow some minors to bypass restrictions. The training data and algorithm biases may also affect recognition performance across different regions and demographics.

Privacy Considerations

Although the system does not require direct identity verification, analyzing account behavior involves data collection. OpenAI needs to clearly explain how this behavioral data is used, stored, and protected.

Summary

OpenAI’s age prediction system represents an innovative attempt at safety governance for generative AI platforms. By using AI to identify AI users, it addresses the practical need for minor protection and demonstrates the potential of technology in self-regulation. However, the true value of this system ultimately depends on its accuracy and user acceptance in real-world applications. As global AI regulations tighten, proactive protective measures like this may become standard features for new products.