Regina Barzilay
MIT Professor, AI for Cancer Detection
MIT professor and breast-cancer survivor who turned her diagnosis into Mirai and Sybil — AI models that predict cancer years early — and co-discovered the antibiotic halicin with deep learning.
In her words
I cannot imagine healthcare without a human physician in the loop.
A lot of low-level tasks will be delegated to AI, but the ultimate responsibility for clinical decisions will remain with physicians.
Biography
Regina Barzilay (born 1970 in Chișinău, then Soviet Moldova, and raised in Israel) made her name at MIT in natural language processing — including deciphering the dead language Ugaritic — and won a MacArthur fellowship in 2017. Her research changed course after a 2014 breast-cancer diagnosis, when she discovered that her own earlier mammograms held patterns no radiologist had flagged.
Teaching machines to see cancer coming
As AI faculty lead of MIT’s Jameel Clinic, Barzilay built Mirai, a deep-learning model that reads mammograms to predict breast-cancer risk up to five years out, and Sybil, which does the same for lung cancer from CT scans — tools validated across hospitals on several continents. In 2020 her team and Jim Collins used deep learning to discover halicin, the first new antibiotic candidate found by AI, and she won the inaugural $1 million AAAI Squirrel AI Award for AI benefiting humanity.
Her advocacy is grounded in patient experience rather than industry interest: she argues AI’s promise in medicine is finding what human eyes cannot — she told PBS the risk patterns are “so subtle” that unaided assertion is impossible — while insisting that models be validated across populations and that physicians retain final responsibility. Her frustration is aimed at how slowly hospitals and regulators adopt tools that could be saving lives now.
Where they stand in the war
Who backs them up
Sources & further reading
Canonical record: https://battlelines.ai/topic/regina-barzilay