OpenAI introduces new safety measures for teen users on ChatGPT

OpenAI is stepping up efforts to safeguard younger audiences on ChatGPT, unveiling a series of protective features designed specifically for teens. The company’s latest initiative, announced this week, highlights a shift toward prioritizing safety and wellbeing over absolute privacy and unrestricted use, marking an important evolution in how artificial intelligence platforms interact with underage users.

At the heart of the new plan is a system that attempts to determine whether someone using ChatGPT is above or below the age of 18. This is not a simple matter, OpenAI admitted, as accurately predicting age in a digital environment presents a host of technical and ethical challenges. When the system cannot determine a user’s age with confidence, the company says it will err on the side of caution, automatically directing the user toward an experience designed for those under 18.

This teen-friendly version of ChatGPT will exclude access to graphic sexual material and other age-inappropriate content. In extreme cases, where a young person appears to be in acute emotional distress, the company has indicated that escalation to law enforcement could be considered. Such a stance reflects OpenAI’s declared principle: when safety and privacy come into conflict, teen safety will take precedence.

Alongside these automated protections, OpenAI plans to roll out a new set of parental controls by the end of September. Parents of teens aged 13 and older will be able to link their own account to their child’s ChatGPT profile through a simple email invitation. Once connected, parents will gain access to a range of settings that let them shape how their teen engages with the AI.

Among these controls is the ability to disable specific features, such as chat history or memory, limiting the information ChatGPT retains across sessions. Parents who choose to activate alerts will receive notifications if the system detects signs that their child may be experiencing distress during a conversation. In addition, a blackout hours feature will allow parents to set times when ChatGPT cannot be accessed, offering an extra layer of control over usage patterns.