OpenAI hires global physicians to vet and improve ChatGPT's medical guidance

OpenAI is building a worldwide panel of doctors to audit the health information produced by ChatGPT, aiming to reduce misinformation and prevent harm. More than 300 million users seek medical advice from the chatbot each week, prompting the company to contract hundreds of physicians who speak 49 languages and operate from places such as Kenya, Nepal and Brazil. Led by Rebecca Soskin Hicks, the team evaluates example conversations, identifies ambiguous or risky replies, and reports shortcomings to OpenAI's research group for further training.

The initiative follows a lawsuit filed by a pastor who claimed the GPT-4o model misdiagnosed his lung condition. OpenAI's head of health products, Ashley Alexander, stresses that the service is not intended for diagnosis or treatment, and the company continues to develop safety measures while navigating an uncertain legal landscape.