United Kingdom, May 12 -- OpenAI has launched a new ChatGPT safety feature designed to alert trusted friends or family members when users appear to be experiencing a serious mental health crisis.

The opt-in feature, called Trusted Contacts, allows users to nominate someone who can be notified if the AI chatbot detects conversations linked to self-harm or suicide risk.

The move comes amid growing scrutiny over the impact artificial intelligence tools may have on vulnerable users, including concerns they could contribute to psychosis, mania or emotional distress.

OpenAI revealed last year that around 0.07 per cent of regular ChatGPT users showed signs of "mental health emergencies related to psychosis or mania".

With roughly 900 million...