AI can identify intimate partner violence years before people disclose it, but is that safe?
MIT researchers built an AI that detects intimate partner violence years before disclosure. The question: should it?
Why it matters
This is a critical AI ethics case study on predictive power vs. privacy and consent. Leaders need to understand how detection capability creates governance dilemmas — flagging risk without explicit disclosure consent could expose patients, yet withholding alerts raises liability questions. It's the kind of AI deployment decision that will define corporate governance frameworks in healthcare.
The key facts
4 to knowMIT and Mass General Brigham researchers developed AI model for intimate partner violence detection
Model can identify IPV risk years before patient disclosure from medical records
Core tension: predictive capability vs. patient privacy and consent
Published June 2026 — emerging healthcare AI ethics case
Go to the source
ZDNet AIzdnet.com
Publisher excerpt: Researchers at MIT and Mass General Brigham have built an AI model that can flag intimate partner violence risk in patients from their medical records.