The Guardian · AI
Devi Sridhar questions reliability of facial AI cancer survival predictions
Researchers at Mass General Brigham developed FAHR-Face, trained on more than 40 million facial images, to assess biological age and predict cancer survival. A survival-focused model used nearly 35,000 photos of cancer patients. Devi Sridhar describes the research as promising but questions reliable predictions for individuals, arguing that many factors have been overlooked.
What to know
- 01
FAHR-FaceAge estimates biological age using facial features including the forehead, wrinkles between the nose and mouth, and temples.
- 02
FAHR-FaceSurvival focuses more closely on the areas below the eyes and the bridge of the nose, using nearly 35,000 cancer patient photos for training.
Why it matters
The research extends doctors’ longstanding use of facial appearance to assess health, while the reliability of individual predictions remains a concern raised by the author.
AI-assisted brief. Read the source for full details. Editorial & sourcing policy
