Large language models for infectious diseases require evidence generation and regulation

Large language models (LLMs) offer significant potential in healthcare, especially in the Australian infectious diseases (ID) context, where a great deal of information must be gathered and synthesised. To maximise benefits, the use of evidence-based medicine principles, robust trials, thorough regulatory frameworks and timely guidelines statements are necessary. Additionally, proactive strategies utilising artificial intelligence architectures such as retrieval-augmented generation can help minimise risks, while optimising the benefits of LLM in ID.