Why are one‑in‑four female collegiate runners sidelined by the same pelvic stress injury while their male teammates stay healthy—and why do women still wait years for diagnoses like endometriosis? In this candid lightning talk, MIT machine‑learning master’s student and former varsity distance runner Lily shares how her own back‑to‑back pelvic fractures exposed a systemic blind spot: less than 1 percent of global R&D targets female‑specific conditions, and women are diagnosed four years later than men across 770 diseases. Lily spotlights MIT’s new Female Medicine through Machine Learning (FEML) initiative, showing how AI models trained on inclusive data can: Screen for endometriosis non‑invasively using multi‑omics and symptom data Detect bias in state‑of‑the‑art chest‑X‑ray algorithms that miss Black, Hispanic, and low‑income women Mine wearables and patient‑reported data for personalized insights at scale She then demos the prototype she’s building with classmates and FEML: a private, AI‑powered chat tool that answers birth‑control questions with evidence‑based, personalized guidance—no Reddit rabbit holes required. Lily call to action is clear: “AI systems mirror what—and who—we value.” With the right datasets, validation standards, and equity‑first mindset, we can stop asking women to wait and start delivering the answers their health demands today. Lily Janjigian is working in the Earth Intelligence Lab, conducting computer vision research for agriculture mapping. Lily was a member of the varsity cross country and track & field teams throughout her undergrad, and is a huge fan of distance running. This talk was given at a TEDx event using the TED conference format but independently organized by a local community. Learn more at https://www.ted.com/tedx