Healthcare AI Blog

Training Physicians and Algorithms in Dermatology Diversity
There's a long-standing challenge in dermatology: Textbooks, databases, journals and lectures are largely bereft of images that feature darker skin.
Read here
Ensuring the Fairness of Algorithms that Predict Patient Disease Risk
Decision-support tools for helping physicians follow clinical guidelines are increasingly using artificial intelligence, highlighting the need to remove bias from underlying algorithms.
Read here
Healthcare Algorithms Don’t Always Need to Be Generalizable
AIMI co-director, Nigam Shah, questions the need for generalizable models and proposes instead sharing recipes for creating useful local models.
Read here
How Do We Ensure that Healthcare AI is Useful?
In healthcare, predictive models need to be more than good predictors. Stanford scholars suggest a framework for determining a model’s worth.
Read here
New AI-Driven Algorithm Can Detect Autism in Brain “Fingerprints”
Led by AIMI faculty Kaustubh Supekar, Stanford scholars have created an algorithm that uses functional MRI scans to find patterns of neural activity in the brain that indicate autism.
Read here
Trust is AI’s Most Critical Contribution to Health Care
AI can reveal remarkable medical insights, but only if patients and doctors have faith in it. Thus, trust has become AI’s singular goal, says Stanford's James Zou.
Read here
Deploying AI in Healthcare: Separating the Hype from the Helpful
AIMI Co-Director, Nigam Shah, assesses the state of AI in healthcare and encourages executives to think beyond the model.
Read here
Broadening the Use of Quantitative MRI, a New Approach to Diagnostics
A promising technology is held back by lack of quality data, but with a newly released dataset, Stanford researchers are about to set it free.
Read here
“Flying in the Dark”: Hospital AI Tools Aren’t Well Documented
A new study reveals models aren’t reporting enough, leaving users blind to potential model errors such as flawed training data and calibration drift.
Read here
The Open-Source Movement Comes to Medical Datasets
Hoping to spur crowd-sourced AI applications in health care, Stanford’s AIMI center is expanding its free repository of datasets for researchers around the world.
Read here
De-Identifying Medical Patient Data Doesn’t Protect Our Privacy
AIMI co-director, Nigam Shah, makes the case that de-identifying health records used for research doesn’t offer anonymity and hinders the learning health system.
Read here
Agile NLP for Clinical Text: COVID-19 and Beyond
With Trove, weakly supervised NLP of clinical text is fast, adaptive, shareable, and high performing.
Read here
Should AI Models Be Explainable? That depends.
AIMI Co-Director, Nigam Shah, advocates for clarity about the different types of interpretability and the contexts in which it is useful.
Read here
When Algorithmic Fairness Fixes Fail: The Case for Keeping Humans in the Loop
Attempts to fix clinical prediction algorithms to make them fair also make them less accurate.
Read here
Algorithm helps detect heart abnormalities
A Stanford AIMI-led team of researchers is using artificial intelligence to detect abnormalities in the heart through an algorithm that assesses the rate at which the heart pumps blood.
Read here
AI Rivals Radiologist-level X-ray Screening for Certain Lung Diseases
In a matter of seconds, a new algorithm read chest X-rays for several possible maladies, performing with about the same or better accuracy than doctors.
Read here