Can one photo of your face reveal how ill you are?

Key Points:

• US researchers have developed an AI tool called 'FaceSurvival' that can estimate a cancer patient's survival prospects by analyzing features around the eyes and nose in a single facial photograph, based on training with millions of healthy faces and 35,000 cancer patient photos.
• The AI aims to assist doctors in making personalized treatment decisions by providing a quantitative, consistent, and scalable health assessment, but it is not intended to replace physician judgment and is still far from clinical use.
• Another tool, 'FaceAge,' predicts biological age by analyzing skin features such as nasolabial folds, marionette lines, and wrinkles, with lifestyle factors like smoking and alcohol use influencing the results; it has shown stability across various real-world photo conditions and some plastic surgery procedures.
• Researchers emphasize the need for more diverse data, especially from different ethnic groups, to validate the tools' accuracy, and they are inviting the public to participate by uploading photos for personalized biological age estimates.
• Experts caution that while the tools show promise as proof of concept, they require further testing against clinical judgments and across varied patient populations before they can be considered for medical use.