What Next ChatGPT, MD
17 snips
Jan 18, 2026 
Brittany Trang, a health tech reporter at Stat News, dives into the transformative role of AI in healthcare. She discusses how AI scribes are alleviating burdens during clinical visits and explores the rise of patient-facing chatbots like ChatGPT Health. Brittany highlights the risks of AI giving misleading clinical advice, emphasizing the importance of nuance in medical judgments. The conversation also touches on the benefits of AI in predictive modeling and reducing administrative tasks, while critiquing tech companies that oversell AI capabilities in health.
AI Snips
Chapters
Transcript
Episode notes
AI Scribers Are Becoming Routine
- AI scribes are already writing clinical notes so doctors don't have to type during visits.
- Large language models (LLMs) are moving beyond scribing into conversational healthcare tools that ingest patient records and devices.
Health Chatbots Aggregate Many Data Streams
- ChatGPT Health and Claude for Health let users upload records and connect health data sources for richer context.
- Companies promise not to train foundation models on uploaded health data, but fine-tuning and cross-chat memory remain unclear.
LLMs Still Produce Dangerous Errors
- LLMs make factual errors and can hallucinate, producing potentially harmful medical advice.
- In tests, chatbots gave potentially harmful advice roughly one in five times, showing real clinical risk.