Stanford University Using Language Analysis for Crisis Hotlines

Many people around the world suffer from different mental health disorders. Among those disorders is depression and anxiety which are two diseases that can affect how a person functions on a day to day basis with relationships, academics, and work life. Crisis hotlines have been put in place to help people suffering with these mental health issues so that they can talk through any bad thoughts they may be having.
There has been a recent emergence in crisis hotlines that can be reached via text. Now instead of calling the hotline and talking to a stranger, people with mental health issues can text throughout the day about how they may be feeling. Graduate students at Stanford University have analyzed hundreds of thousands of texts from the thousands of text conversations between  people with mental health disorders and the counselors at the crisis hotline. They were looking to find a way to determine whether a textual conversation had been effective or not. The researchers looked at natural language analysis to determine whether a certain way of texting improved the way the person felt after the conversation.

The researchers found that the successful conversations all had five stages to them (Introduction, problem setting, problem exploration, problem solving, and wrap up of the conversation) that were all marked by key words. The researchers are hoping that by analyzing the language of the crisis counselors, they may be able to generate an automated counseling system to increase the amount of people that can be helped. They are also hoping to use artificial intelligence to make the automated counseling seem more human-like and approachable for a person who has a mental health disorder.