In a recent study led by the University of Cambridge, researchers found that the AI model GPT-4 surpasses the diagnostic capabilities of non-specialist doctors in evaluating eye problems. The study, published in the journal PLOS Digital Health, compared the performance of GPT-4 with doctors at various career stages, including junior doctors and expert eye specialists.
During the study, participants were presented with 87 patient scenarios involving different eye issues and asked to provide diagnoses or treatment recommendations. GPT-4 consistently outperformed unspecialised junior doctors and achieved similar scores to both trainee and expert eye doctors.
Despite these promising results, researchers emphasise that AI models like GPT-4 are not intended to replace healthcare professionals but rather to complement their expertise. They suggest that such models could be valuable in aiding healthcare workflows, particularly in tasks like triaging patients with eye problems.
Dr. Arun Thirunavukarasu, lead author of the study, highlights the potential for AI to assist general practitioners who may require prompt advice from eye specialists. Additionally, he notes the importance of characterising the capabilities and limitations of commercially available models, as patients may already be utilizing them for medical advice.
The study’s findings shed light on the evolving role of AI in healthcare and underscore the need for further research and development in this field. With advancements in AI technology, the potential for improving patient care and enhancing healthcare workflows continues to grow.