How suicide prevention app alerts you to signs of distress on Twitter

Tapping into a public stream. Indiana Public Media, CC BY-NC

For more or less any topic we can think of, social media has great potential for both good and harm. One example is suicidal distress. It is increasingly common for people to express despairing and suicidal thoughts on social media sites. On the one hand this is to be welcomed, as more public expression of such feelings could lead to a reduction in stigma. However, the public airing of statements and images associated with suicide and self-harm could also feed into a normalisation of these behaviours whereby they become seen as less extreme responses.

Normalisation is likely to be one of the mechanisms in the development of suicide clusters. Given that most people posting about emotional distress online are seeking some kind of help, there is also major potential for suicide and self-harm prevention via the internet. However, it is very early days and there is not yet any strong evidence of effective approaches.

With the help of a creative marketing agency, the Samaritans have launched an app called Samaritans Radar. If you sign up for the app, an alert comes into your inbox if key words and phrases used for despair and distress appear in the tweets of anyone you follow on Twitter. It only works with publicly available tweets, so no protected accounts can be included. At the moment the app knows no geographical boundaries so it will work with tweets from anywhere in the world.

The language of distress

Our research in Cardiff, (part of the Cosmos project) provided some of the key words for the app’s algorithm. These were identified via linguistic analysis of specialist websites that people visit to express suicidal thoughts. Some of the words and phrases are explicitly about a threat to life and others are indications of despair, hopelessness and worthlessness.

The app is geared towards younger people and initial testing was done with 300 young people and from this, the language was refined to remove some false positives.

Samaritans Radar is a positive step forward but not a panacea. It is early days for such interventions and there will be teething problems. One challenge is the difficulty of avoiding flippant statements, such as feeling suicidal because of a boring maths lesson.

Creeping surveillance?

A Samaritans Twitter discussion brought up comments from a few Twitter users about creeping surveillance. There needs to be a debate about the balance of protection and civil liberties and not everyone will agree about where to strike the balance.

The key thing to remember is that only publicly available postings are tracked. These are tweets that the app user could have seen anyway but may have got lost in the flow. And for many this would be considered benign surveillance, which is meant to increase human connection through offers of concern and support when people are struggling – it is not about professional intervention but about improving support from friends in difficult times.

The app does not currently work with other social media platforms, but there is room for future development. Facebook is quite different from Twitter as most people have private Facebook accounts only accessible to friends.

One online phenomenon which is challenging is the deliberate posting of suicide and self-harm material within committed communities. This currently happens in Tumblr in particular. These are people who actively want to post material about despair and distress, perhaps with a virtual identity created for that specific purpose and who may not always be so open to offers of help. This kind of social media phenomenon needs a different approach.