OpenAI is rolling out new safety features for under-18 users of ChatGPT, including switching off the human voice response, regular break reminders, and prompts reinforcing that ChatGPT is AI rather than a person.

Teen accounts will become the default for anyone who identifies as a teenager when signing up, or who OpenAI itself identifies as under 18, though the company has not detailed how this identification will work. Under-13s are not permitted to use the product.

The update lands amid growing unease among experts about young people’s relationships with chatbots. Dr Susan Shelmerdine of Great Ormond Street Hospital has compared heavy AI use to ultra-processed food, warning of “ultra-processed minds”.

There is currently no bespoke AI regulation for chatbots in the UK, and the move follows a separate Policy Exchange report claiming most UK universities remain exposed to AI assisted cheating in remote assessments.