With the advancement of artificial intelligence, the European Medicines Agency released " Guiding principles on the use of large language models in regulatory science and for medicines regulatory activities" in 2024, emphasizing the importance of the safe and responsible use of large language models. This guidance covers general ethical considerations, user principles and organizational principles, clearly outlining the potential of large language models in areas such as text processing assistance and data mining, while also warning of risks such as hallucinations, data privacy issues, and biased outputs. It proposes measures including continuous learning, risk monitoring and mechanism-building to address these challenges. Although China has not yet issued similar guidelines, large language models hold potential application prospects in areas such as assisting in the processing of review materials, formulating and revising guidance principles and identifying risks. At the same time, challenges related to decision interpretability, data bias, legal accountability and talent reserves remain. It is recommended that China’s regulatory authorities draw on international experience to construct an application framework tailored to national conditions as soon as possible, thereby promoting the integration of artificial intelligence technology with the field of drug regulation.