Suicide results in approximately 44,965 American deaths each year and is the tenth leading cause of death in the United States, according to the American Foundation for Suicide Prevention. Deaths from cardiovascular disease and other ailments greatly outnumber those caused by suicide, but suicide rates have either remained steady or even increased in certain parts of the U.S., while heart condition casualties have been decreasing.

For science journalist Lydia Denworth, this is a clear sign that current efforts to prevent suicide aren’t working. She believes doctors simply aren’t able to identify potential risk factors and act on them for all of their patients. Denworth tells Cheddar’s Morning Bell that artificial intelligence developed by social scientists could predict who is most at risk so doctors can act more effectively instead of spending time analyzing medical records.

suicide rates vs heart disease and cance
Rates of heart disease and cancer in American men and women from 1980 - 2010 (left). Overall American suicide rates in 2007-2016 (right).

“The problem was needing to look far beyond single risk factors,” she explains. “They’ve brought in machine learning as a way to look at thousands of risk factors at a time, far more quickly and efficiently than a human being can do. This is still very new, it’s only being done experimentally but the idea is that eventually there will be software that will put a red flag up on someone’s records.”

Denworth went on to say that this wouldn’t completely replace human doctors’ role in suicide prevention. Instead of completely automating the process, this A.I. tool would utilize the huge amounts of medical data that is stored electronically to tell medical professionals when to act.

This is a perfect example of machine learning and humans working in tandem to solve a major issue. Once this A.I. system has been fully developed and is tested for accuracy it can be used make a change that humans have not been able to achieve alone.