Facebook has started using a new artificial intelligence (AI) system to identify users showing warning signs of suicide.
The AI has been trained to spot status updates and subsequent comments that could indicate someone was at risk.
The algorithm would immediately send a report to a real reviewer, who could then contact the user with suggestions and resources to help if appropriate.
At the moment, Facebook relies on a human reporting system regarding potential suicides, where friends of users can click a button to tell the company about concerning updates.
But it is hoped more people will be helped by an automated system.
The AI will look out for statuses that suggest the user is sad or in pain, as well as responses from friends expressing concern.
Facebook is also adding suicide prevention tools to Facebook Live after several deaths have been broadcast since the feature was introduced.
This would give viewers the option to report a friend if they are concerned, and gives the reporter resources to help.
The broadcaster at risk will also be given the option to contact a friend, mental health helpline or see tips.
At the moment, this is only being trialed in the United States.
Head of media for the charity Samaritans Lynsey Pollard said: “It’s great that Facebook are showing such commitment to keeping people safe online.
“Samaritans was involved in the consultation for their UK safety features and we would hope to have the same involvement if they bring this latest development to the UK.”