Multidisciplinary blinded randomized expert evaluation of large language models for clinical diagnosis and management

Key Points

• To assess the clinical application of large language models and identify associated risks.
• Conducted a blinded randomized evaluation by multidisciplinary experts.
• Developed an evaluation framework for large language models.
• Validated the framework through expert consensus.
• Identified substantial potential of large language models in medicine.
• Highlighted the serious risks associated with unmonitored clinical use.
• Emphasized the need for continuous oversight by human experts.