Grounding (AI)

Grounding in AI refers to tethering a model's outputs to verifiable, real-world sources — retrieved documents, database records, structured data, or tool results — rather than allowing it to generate from parametric memory alone. Grounded systems cite their sources and are auditable.

Grounding is the primary enterprise defense against hallucination. A grounded RAG system retrieves the three most relevant document passages and instructs the model to answer only from those passages, producing a citation. Grounding reduces but does not eliminate hallucination — models can misinterpret or selectively quote retrieved content. Regular evaluation against a golden dataset is required to measure grounding effectiveness in production.