Algorithms Help Spot Possible Suicidal Intent Among Veterans’ Social Posts
A new social media platform that is designed to connect both active and retired US Military members has now been outfitted with a new custom machine learning model that can allegedly predict signs of potential self-harm or suicidal intent. The algorithm assesses posts made on the platform and identifies those that signify the user is contemplating or engaging in self-harm. The Veterans Affairs Department, Harvard University, and Amazon Web Services formed a partnership with the RallyPoint military social media platform to develop the machine learning program.
The technological solution quickly surfaces sensitive public posts and creates an efficient way to boost online suicide intervention. RallyPoint has been around since 2012 and boasts nearly 2 million members, including active service, veterans, and military families. The platform enables these community members to chat about topics involving military life, ask questions, share stories, and other functions. The new program is an attempt to lower the veteran suicide rate, which is currently 50% higher than that of non-veteran American adults.