What if artificial intelligence and natural language tech could diagnose suicidal tendencies early enough to save lives? Don Wright was working on a solution that October night in 2017. He played ping-pong with his 28-year-old son, Justin, and all seemed well. They had discussed the future; it seemed bright. But in the morning light, a friend called to say Justin was missing. Wright had a terrible feeling. “I think he’s dead,” he told his girlfriend, moments before arriving at his son’s apartment. Many parents of a child dealing with depression are familiar with Wright’s feeling of incoming tragedy. Suicide is the second-leading cause of death in the U.S. for people aged 10 to 34, and more than 47,000 Americans killed themselves in 2017, according to the American Association of Suicidology. Approximately one in five teenagers experience some form of severe mental disorder, yet despite the prevalence of mental health issues, it’s still treated as “something mystical,” Wright says. “We don’t say ‘we think he has a tumor, so we’re going to give a drug to treat it.’ We eventually know what’s wrong with somebody in other parts of medicine. We need to get there with mental health.” | 
