American scientists at Boston's Mass General Brigham medical center have developed two artificial intelligence tools that use an ordinary facial photo to determine a person's biological age and the risks associated with cancer treatment, infohub.kz reports.

The base model, called FAHR-Face, was trained on more than 40 million portraits. Two programs were then built on it: FAHR-FaceAge, which estimates apparent age, and FAHR-FaceSurvival, which predicts survival in cancer patients.

Study co-author Raymond Mak explained that the new technology will help doctors more accurately assess the body's physical reserve before prescribing aggressive therapy or surgery.

A test on 38,183 patients showed that people who looked at least five years older than their passport age had a 21% higher risk of death. The neural network is also nearly impossible to fool with lighting, facial expressions, makeup or plastic surgery, although it clearly picks up traces of smoking and alcohol use.

The second program rates risk on a scale from zero to one: in the test group, patients with the highest scores died three times more often than those who received the lowest scores.

Interestingly, the algorithms analyze different areas of the face: the forehead, temples and nasolabial folds matter for determining age, while the area under the eyes and the bridge of the nose matter for assessing survival.

The authors note that the programs still require additional clinical trials in different countries, since risk-assessment accuracy varies among people of different races. The researchers have now launched an open project where anyone can upload a photo and find out their biological age.