Glossary
The vocabulary of grounded AI
The words that decide whether an AI answer can be trusted. Each one is defined here in the sense that matters when the answer has to hold up in front of an auditor.
-
Grounding
Grounding is the discipline of making an AI answer only from a verified body of knowledge, cite the source of every claim, and say what it cannot find instead of guessing. This is the document sense of the word, the one an industrial buyer means.
Read the definition → -
Hallucination
A hallucination is a confident, fluent statement from an AI that is not supported by any real source. In grounded systems the dangerous case is subtler: an answer that is sourced, plausible, and still wrong because a condition was left behind.
Read the definition → -
Silent omission
Silent omission is what happens when a retrieval system leaves something out and the answer gives no sign of it. The output looks identical whether a passage was read and set aside or never opened at all.
Read the definition → -
Traceability
Traceability is the property of being able to follow an AI answer back to the exact source it came from. An answer you cannot trace is one you have to re-check by hand, which is the real ceiling on enterprise AI.
Read the definition → -
Provenance
Provenance is the origin and history of a piece of knowledge: where it came from, which revision it is, and in what form it was captured. In regulated work, a fact without its provenance is a fact you cannot rely on.
Read the definition → -
Ground truth
Ground truth is the verified body of knowledge an answer is checked against: the documents an organization has decided are correct and authoritative. It is the thing an AI answer has to be grounded in, and the thing you audit against.
Read the definition → -
Domain-Specific Artificial Intelligence (DSAI)
We call this DSAI: a Domain-Specific Artificial Intelligence, a compact, reliable expert in one domain. Once a domain's knowledge is grounded, being right no longer takes a frontier-scale model. It takes a focused one.
Read the definition →