OpenAI has set out its approach to teen safety on ChatGPT, arguing that keeping under-18s away from AI until adulthood would leave them badly prepared for a defining technology of their generation.
In a blog post published on 16 July, the company said nearly nine in ten teens using ChatGPT do so weekly for learning, information, skill-building or productivity. Its safety framework rests on four commitments: prioritising teen safety even where it conflicts with other goals, encouraging real-world support, treating teens as teens rather than adults, and being transparent about expectations.
The company automatically detects likely under-18 users and applies extra protections designed to limit exposure to violent, self-harming or body-image content, viral challenges, and inappropriate roleplay scenarios. Parents can now activate Study Mode remotely via Parental Controls, and will receive notifications if a linked teen account is deactivated for violent threats or acts of violence, alongside existing self-harm alerts.
On the learning side, OpenAI said 18 million weekly users now engage with its interactive maths and science tools, covering more than 250 topics from integrals to photosynthesis, while a pronunciation feature now supports over 61 languages.
OpenAI has joined the Family Online Safety Institute this year as part of a wider push involving the American Psychological Association, the American Federation of Teachers and Common Sense Media.
