The tragedy of Alzheimer's disease is partly a tragedy of timing. By the time symptoms are obvious enough to diagnose, much of the damage is already done. Alice Tang, 29, is building artificial intelligence tools designed to close that gap, listening for the disease years before it announces itself.
Diagnosis Hidden in Plain Sight
Named to the Forbes 30 Under 30 Science list for 2026, Tang works on using AI to understand and diagnose complex diseases like Alzheimer's. Her approach reframes what a diagnosis even is. Rather than waiting for a single decisive test, she treats a patient's accumulated medical history as a vast, messy dataset in which the earliest signs of dementia may already be recorded, scattered and unnoticed.
The insight is that the clues are often there long before anyone connects them. A pattern of seemingly unrelated conditions, ordered in a particular sequence over years, can carry a signal that no individual doctor's visit would reveal.
Why Complex Diseases Resist Simple Tests
Alzheimer's does not behave like a broken bone. It emerges from a tangle of genetic, metabolic, and vascular factors that unfold over decades. That complexity is exactly what makes machine learning valuable, because algorithms can hold thousands of variables in view at once.
- Longitudinal signals: patterns that only become visible across years of records.
- Sequence matters: the order in which conditions appear, not just their presence.
- Early windows: identifying at-risk patients while intervention could still help.
- Interpretability: ensuring clinicians can trust and understand the model's reasoning.
The Promise and the Peril of Prediction
Predicting disease before it strikes is a double-edged gift. On one side lies the possibility of earlier intervention, lifestyle changes, closer monitoring, and enrollment in trials while treatment could still matter. On the other lies the ethical weight of telling someone they may be headed toward a condition with, as yet, no cure.
Researchers in this space, Tang among them, are acutely aware that a prediction is only as useful as the action it enables. An algorithm that flags risk without a path forward risks becoming a source of dread rather than help. The most thoughtful work in the field ties prediction to intervention.
A Generation Fluent in Both Medicine and Machines
What makes Tang emblematic of her cohort is her fluency across two languages that rarely met a generation ago: clinical medicine and computer science. The Forbes 2026 science class is full of young researchers who move comfortably between the wet lab and the codebase, treating biology as an information problem.
That fluency is reshaping how diseases are studied. Where earlier medicine waited for symptoms and then reacted, this new approach mines the past to anticipate the future. It will not cure Alzheimer's on its own. But by moving the moment of detection earlier, it changes what is possible, giving patients, families, and doctors something the disease has long denied them: time.