Benchmarks & Evaluations

What It Takes to Build a Medical Database

From fragmented patient records to Al-ready insights.

Aug 18 • Sam Ashoo, MD

OpenAI's Uncontrolled Experiment

A sandbox breach, a 52 percent under-triage rate, and two lawsuits all point to the same architecture OpenAI now uses to read patient charts.

Jul 24 • Sam Ashoo, MD

Who Grades the Evidence? Five Products, Five Different Answers

OpenEvidence’s new EvidenceGrade vs UpToDate, DynaMed, Vera Health and Consensus

Jul 14 • Sam Ashoo, MD

When Clinical AI Meets Independent Evaluation

Real-World Testing Shows a Different Picture

Jun 14 • Sam Ashoo, MD

11 Medical AI Tools Read These Xrays

Every One Missed The Pneumothorax

Jun 9 • Sam Ashoo, MD

Can Medical AI Read an ECG?

Last week, I published the results of a challenging but routine case posed to 6 AI models.

Jun 2 • Sam Ashoo, MD

5 Medical AI Models Got This Case Wrong. Is Your Favorite One of Them?

The highest-profile tools were not necessarily the safest.

May 28 • Sam Ashoo, MD

OpenEvidence, Doximity, and the FDA

The Gray Zone in Clinical AI

May 15 • Sam Ashoo, MD

AI's Invisible Error

Would We Accept This Anywhere Else in Medicine?

May 7 • Sam Ashoo, MD

Same Model. Different Database. Different Answer.

Rag is the Reason These Ai Tools Feel So Different

May 4 • Sam Ashoo, MD

ChatGPT for Clinicians: What Kind of Tool Is This?

ChatGPT for Clinicians is being positioned as a tool for documentation, clinical questions, and literature review.

May 1 • Sam Ashoo, MD

#nojs-banner { position: fixed; bottom: 0; left: 0; padding: 16px 16px 16px 32px; width: 100%; box-sizing: border-box; background: red; color: white; font-family: -apple-system, "Segoe UI", Roboto, Helvetica, Arial, sans-serif, "Apple Color Emoji", "Segoe UI Emoji", "Segoe UI Symbol"; font-size: 13px; line-height: 13px; } #nojs-banner a { color: inherit; text-decoration: underline; } This site requires JavaScript to run correctly. Please turn on JavaScript or unblock scripts