•Eric Topol highlights four AI models—Delphi 2M, APOLLO, ALADYNOULLI, and AURORA—that forecast personal health risks up to 20 years ahead, including condition onset timing and magnitude, using large-scale data from UK Biobank, Mass General Brigham, and All of Us.
•The models incorporate varied inputs such as electronic health records, lifestyle factors, polygenic risk scores, 25 billion clinical events, and multi-omics to enable personalized predictions.
•They support “what if” simulations to model how lifestyle changes or medications could alter an individual’s future health trajectory, advancing precision prevention.
This is huge because it makes prevention personal.
Instead of generic “eat better, exercise more”, AI can say:
“Keep current habits = 60% diabetes risk by 65.
Lose 7% body weight + 150min walking = 22% risk.”
That specificity is what actually gets people to act.