A team of researchers at Mass General Brigham says it could at least provide a set of powerful new clues.
The technology represents a move in facial recognition towards prognostic medicine and offers an intriguing glimpse into the intersection of digital health, consumer tech and oncology.
The study found that lifestyle factors, especially smoking, heavy alcohol use and drug use, heavily skew biological age older than chronological age.
Combining the biological age metric with the direct survival score yielded an even more accurate model fit, the researchers state.
The AI identifies statistical patterns linked to poorer health outcomes, but it does not explain the biological mechanisms driving them.