A recent study led by the University of Cambridge and the University of Birmingham, along with collaborators worldwide, has shown that the AI model GPT-4 has the potential to outperform non-specialist doctors in evaluating eye issues and providing guidance. The study, published in the journal PLOS Digital Health, tested GPT-4 against various medical professionals, including junior doctors and expert eye doctors, in diagnosing and providing treatment advice for different eye conditions. GPT-4 demonstrated superior performance compared to junior doctors, aligning closely with both trainee and expert eye doctors.
While large language models like GPT-4 are not meant to replace healthcare professionals, they can enhance healthcare by offering guidance on eye-related matters and providing diagnoses. These models could be particularly useful in triaging patients or in situations where access to specialized healthcare providers is limited. The study highlights the potential for AI to complement and enhance modern medicine, especially in the field of eye disorders.
GPT-4, along with other models like GPT-3.5, PaLM 2, and LLaMA 2, are extensively trained using vast datasets of words from various sources. GPT-4, in particular, has shown to provide more precise and accurate responses compared to other models. The field of large language models powered by AI is rapidly evolving, with newer and more sophisticated models being introduced that could potentially rival the expertise of professional eye doctors.
Source link
Source link: https://www.birmingham.ac.uk/news/2024/artificial-intelligence-outperforms-physicians-in-the-precise-diagnosis-of-eye-disorders
GIPHY App Key not set. Please check settings