Artificial intelligence can predict the biological age of self-photographs

A new model of artificial intelligence, called Face Age, estimates how old a person looks compared to her chronological age. Manufacturers say their tool could help doctors decide on the best way to treat diseases such as cancer.
« When a doctor treats a cancer patient, one of the first things he does is try to assess how well the individual feels, » said Hugo Ert, director of the Medicine General Intelligence Intelligence Program at Massachusetts General Hospital, Brigam.
According to him, it is often a very subjective assessment, but can affect many future treatment decisions, including how aggressive or intense their treatment plan should be. For example, doctors may decide that a patient who looks younger and better responding to his age can better tolerate aggressive treatment and eventually live longer than a patient who looks older and more, even if both are the same age. Face Age could facilitate that decision by converting subjective assessments of doctors to a quantitative measure, the study’s authors wrote, and by quantification of biological age, the model could offer another data point that would help doctors decide which treatment to recommend.
ERTS and his colleagues trained the model of more than 58,000 photos of people aged 60 and older who were believed to be in average health for their age at the time the photo was taken. In this training set, researchers assessed the chronological age of the models and assumed that people’s biological age was similar, although scientists noticed that this assumption was not true in any case. The team then used Chase Age to predict the age of more than 6,000 people with cancer. Cancer patients looked, on average, about five years older than their chronological age.
However, Chase Age is not yet ready for use in hospitals or medical practices. Further research with larger and more representative training is needed to understand how these factors influence assessments.