AI Figured out the Word People Text When Their Suicide Risk Is High

March 24, 2017

News Type:  Weekly Spark, Weekly Spark News


Crisis Text Line is using artificial intelligence (AI) to identify words associated with suicide risk and improve the effectiveness of its services. The crisis support line has developed an algorithm to analyze data from 22 million text messages and determine the words people use when they are at high risk for suicide. Based on its findings, the non-profit has incorporated 9,000 words and word combinations into its algorithm, which triages messages based on risk level. According to Crisis Text Line CEO Nancy Lublin, this approach allows the service to quickly and accurately identify users at highest risk and provide them with the crisis support and resources they need ahead of low-risk users. Since the algorithm was introduced, volunteer response time to high-risk texts has decreased from under 2 minutes to 39 seconds. Julie Cerel, a clinical psychologist and president-elect of the American Association of Suicidology, said that this new approach indicates that “we are finally listening to people who are suicidal and using what they’re telling us to figure out how to help them.”

Spark Extra! Learn more about Crisis Text Line and Crisis Trends data.