Technology
RAG (Retrieval-Augmented Generation)
Pulling supporting documents into the AI's context before it answers.
RAG (Retrieval-Augmented Generation) is the pattern of grounding an AI response in retrieved source documents. In healthcare AI it's the technique behind citation-grounded answer engines (e.g. OpenEvidence) — the model isn't free-associating; it's synthesising from retrieved literature and showing the citations.
For scribes, the same technique can be used to keep generated notes anchored to what was actually said, rather than letting the model paraphrase freely.
See also
- AI hallucination — When an AI model writes something confidently into a note that was never said.