Meet OpenAI's doctor-whisperer making ChatGPT better at talking health
OpenAI enlisted over 260 global physicians to improve ChatGPT's health advice amidst legal challenges and a high-stakes healthcare push.
TL;DR
- Over 300 million people use ChatGPT weekly for health advice.
- OpenAI has hired hundreds of practicing physicians globally to review ChatGPT conversations and identify areas for improvement.
- Rebecca Soskin Hicks leads the team of physicians, working to enhance AI's safety and accuracy in healthcare.
- Physicians evaluate conversations for potential user confusion or danger, with a focus on ambiguous and complex scenarios.
- Despite a lawsuit alleging misdiagnosis, OpenAI is continuing to develop its healthcare AI capabilities.
- OpenAI states that ChatGPT is not intended for diagnosis or treatment, but rather as a tool to aid information seeking.
- The company aims to integrate health records and has developed benchmarks for testing AI's health and mental health responses.
- The goal is for physicians' feedback and performance tests to establish a growing body of evidence for AI in health.