Applications and Future Prospects of Large Language Models in Healthcare

This paper provides a systematic review of the development approaches, model types, and current applications of large language models (LLMs) in the healthcare domain. It highlights the unique and increasingly complex ethical and safety challenges these models face in healthcare contexts. Compared with other artificial intelligence technologies, LLMs exhibit deeper issues related to data security and privacy protection, algorithmic bias and fairness, the reconfiguration of doctor–patient relationships, ambiguity in the attribution of responsibility, and ethical risks arising from cross-domain transfer. The practical implementation of LLMs in healthcare requires not only continuous technological advancement to improve the accuracy and interpretability of generated content, but also a careful consideration of their potential ethical risks and societal implications. The integration of LLMs into clinical workflows should be guided by the four fundamental ethical principles of autonomy, beneficence, non-maleficence, and justice.