US Veterans Affairs suicide prevention algorithm favours white men
A US Department of Veterans Affairs (VA) AI programme called REACH VET that prevents suicide amongst military veterans has been found to prioritise white men and overlook survivors of sexual violence . According to a Markup investigation, t he algorithm considers 61 variables and gives preference to veterans who are “divorced and male” and “widowed and male”, but does not prioritise female veterans . Notably, it does not take into account military sexual trauma and intimate partner violence, both of which are linked to elevated suicide risk among female veterans . Recent government data shows a 24 percent rise in the suicide rate among female veterans between 2020 and 2021, which is four times the increase among male veterans during the same period . Despite this, the VA’s suicide prevention algorithm continued to prioritise men . The VA claims that its machine learning model, REACH VET, flags 6,700 veterans a month for extra help and results in a significant reduction in suicide attempts by those veterans over the following six months . However, the algorithm’s bias towards white men and its neglect of survivors of sexual violence raised concerns about its effectiveness and fairness . System 🤖 REACH VET Documents 📃 US Department of Veterans Affairs (2018). REACH VET Overview (pdf) US Department of Veterans Affairs (2017). VA REACH VET Initiative Helps Save Veterans Lives: Program Signals When More Help Is Needed for At-risk Veterans Operator: Developer: US Department of Veterans Affairs Country: USA Sector: Govt - welfare; Govt - defence Purpose: Prevent suicide Technology: Machine learning Issue: A ccuracy/reliability; Bias/discrim i nation - race, ethnicity, gender Investigations, assessments, audits 🧐 The Markup . V.A. Uses a Suicide Prevention Algorithm To Decide Who Gets Extra Help. It Favors White Men
- Date it happened
- 2024-05-01
- Organisation involved
- US Department of Veterans Affairs
- Product, system or model
- REACH VET
This incident was imported from AIAAIC and is used under CC BY-SA 4.0. Our additions to it — the structured fields, the translation, the checks against other reports — are published under the same licence.
This is a record of what was reported, not a finding that anyone broke the law. If it names your organisation and you believe it is wrong, the corrections process is free and open to everyone.