<ul><li>A study evaluated Large Language Models (LLMs) for generating cancer-related information to support patient understanding.</li><li>General-purpose LLMs produced higher linguistic quality and affectiveness, while medical LLMs were better at communication accessibility.</li><li>However, medical LLMs showed higher potential harm, toxicity, and bias, impacting their safety and trustworthiness.</li><li>The study emphasizes the importance of intentional model design to improve safety, mitigate harm and bias, and enhance the quality of AI-generated health content.</li></ul>

Large Language Models for Cancer Communication: Evaluating Linguistic Quality, Safety, and Accessibility in Generative AI

Discover more