Assessing the role of Large Language Models in Mild Cognitive Impairment Management: A Bilingual Analysis Comparing ChatGPT, Gemini, and Kimi

This study demonstrates LLMs' proficiency in symptom-related inquiries and stronger alignment with healthcare professionals' operational needs versus care partners' accuracy priorities. English responses outperformed Chinese due to corpus disparities, highlighting needs for language-specific optimization. Findings underscore LLMs' clinical potential, urging enriched Chinese medical corpora and specialized model development to address care partners' unmet requirements.