Practical AI in Healthcare S1, E13 - Part 2 of 2: Dr. Yin Ho's discussion of her new book, Rushing Headlong: Health IT’s Legacy and the Road to Responsible AI
Nov 23, 2025
Dr. Yin Ho, a Health IT strategist and author, brings sharp perspective on healthcare technology and responsible AI. She examines risks of large LLMs, the case for smaller domain-specific models, and why data quality and provenance matter. Topics include decision support versus control, ambient scribing pitfalls, and strategies to avoid repeating past digital missteps.
AI Snips
Chapters
Books
Transcript
Episode notes
Decision Support Versus Decision Control
- The human choice to let systems act determines whether AI stays as decision support or becomes decision control.
- S. Yin Ho warns automation complacency can shift responsibility away from clinicians over time.
Abstract Clinical Meaning Before Training
- Training on messy EMR notes risks encoding poor-quality signals into models.
- Ho recommends abstracting and curating clinical meaning before using data for training.
Start With High‑Value Questions For Abstraction
- Focus on common research questions and diseases to prioritize abstraction efforts.
- Build labeled, structured inputs from unstructured notes before training targeted models.
