AI glossary · Basic · Language models · Safety & alignment · Healthcare & bio
Medical LLM
A medical LLM is a large language model adapted to medical data such as medical knowledge and clinical records, to help with answering questions, writing documents and summarizing information in clinical settings.
About this term
- Level
- Basic · Technical terms you often meet in AI news
- Category
- Industry AI
- Industries
- Healthcare & bio
- Added
- 2026-10-10
How it works
The basic foundation is the same as a general LLM. A model pretrained on a large amount of text is either fine-tuned with additional training on medical textbooks, papers, clinical guidelines and de-identified clinical records, or used with RAG, which retrieves trustworthy medical sources and adds them to the question. As a result, it better understands medical terminology and clinical document formats and answers in the language medical staff use. Recently, more multimodal models that also handle images and test values are appearing.
Why it matters
Medical staff spend a large share of their working time writing records and checking information. Medical LLMs can reduce this burden by summarizing clinical conversations into draft records, condensing long patient histories into an at-a-glance overview, or quickly finding the latest guidelines. For patients, they can serve as helpers that explain difficult test results or medication instructions in plain language. Used well, they can help medical staff spend more time face-to-face with patients.
Things to know
LLMs can hallucinate, making up content that sounds plausible but is wrong, and in medicine such mistakes directly affect patient safety. That is why a human-in-the-loop structure, where answers show their sources and medical staff do the final check, is considered essential. Clinical data is highly sensitive personal information, so de-identification and access control are important. Depending on the country, features used directly in diagnosis or treatment decisions may be subject to medical device regulation (SaMD).
Checks before adopting a medical LLM
Safety
- Does it show the sources for its answers?
- Do medical staff do the final check?
Data & regulation
- Has clinical data been de-identified?
- Is access controlled?
- Has it been checked whether medical device regulation applies?
Example
A use commonly tried in hospitals is "drafting clinical notes." The conversation in the exam room is converted to text through speech recognition, and the medical LLM summarizes it into a set format such as symptoms, findings and plan to create a draft for the electronic health record (EHR). The doctor reads and edits this draft before saving it. Other uses include drafting replies to patient inquiries and rewriting discharge instructions in plain language.