Improving health intelligence in ChatGPT

### TL;DR OpenAI has assembled a team of practicing doctors to assess and improve ChatGPT’s responses in health-related topics based on real clinical experience. The goal is to enhance accuracy and communication, ultimately making medical knowledge more accessible. ## Why Doctors Are Needed for AI Health Models We live in an incredibly complex healthcare system. This technology can greatly improve our way of life, and if done right, it holds enormous potential. Doctors choose medicine because they want to care for others. At age 13, I saw my father change in ways I never expected, which truly ignited my passion to become a doctor. Both my grandfather and uncle were diagnosed with schizophrenia, leading me to choose psychiatry as my specialty. I decided to become a pediatrician to express that care, and what I’m contributing to the AI team now is a continuation of that desire. ## The Unique Composition of the Doctor Team I’ve been a pediatrician for 16 years. I lead a team of hundreds of doctors who evaluate OpenAI’s health models to ensure they are useful and safe. Our doctors bring real medical experience into ChatGPT’s health models. We are all practicing physicians who see patients every day—that’s what makes this team so unique. ## The Core of Evaluation Work A lot of our work involves evaluating example conversations. We focus on the accuracy of responses and also assess the impact a response can have on a person. How something is expressed is often as important as what is expressed. It used to be easier to spot what needed improvement in a response; now we can see the results of our work in real time, which is really cool. It has improved a lot, and I’m genuinely proud of this team because they are a key reason users can trust the information they receive. ## Hopes for the Future My hope for the future isn’t just about what AI or ChatGPT can do, but about making it accessible to everyone around the world. I’m a hematologist and transfusion medicine specialist—a very niche field. The promise of AI for me is to democratize this knowledge, to bring it to places I can’t go, but perhaps an AI model can. Simply lowering the barrier to accessing information on health topics has an enormous impact. We have all taken an oath to be responsible to patients. I believe the core of all this is truly empowering people in their health. As a doctor, I’m not afraid of this; instead, I hope for more such opportunities. I love seeing patients in my clinic, about 15 a day; through this work, I can impact millions of lives—that means a lot to me. Source: https://www.youtube.com/watch?v=UxY8zJKRrHU