|
Briefing
|
- AI-powered Mental Illness Detection – Scientists at the University of Southern California, Carnegie Mellon University, and Cincinnati Children's Hospital Medical Center used machine learning to distinguish suicidal from depressed patients, discovering smile as good indicator to identify suicidal patients
- Distinguishing Smiles – Identified patients with non-Duchenne smile, kind of smile without contraction of muscles surrounding eyes, are more likely to exhibit suicidal tendencies
- Experimentation – Administered study on 379 subjects (i.e. 123 control, 126 mentally ill and 130 suicidal patients), conducting interviews and recording smiles and other facial behaviors, then fed data into machine-learning algorithm that looked for correlations
- Potential Implications – Commercial use of algorithm may lead to discriminative behavior, negatively affecting person's ability to purchase life insurance, receive promotions, or get employed
- Next Steps – Wants to make work available for clinical purposes to help doctors determine and differentiate suicidal and depressed patients
|
|
Accelerator
|
|
|
Sector
|
Healthcare/Health Sciences
|
|
Organization
|
Carnegie Mellon University, Cincinnati Children's Hospital Medical Center, University of Southern California
|
|
Source
|
|
|
Original Publication Date
|
June 15, 2017
|
