Decoder. plain-English AI glossary

Grounding

● Core

Tying a model’s answers to real, checkable sources instead of letting it free-associate.

Think of it like

Like a journalist who cites documents rather than writing “trust me.”

Example

A support bot grounded in the actual help docs quotes the real refund window instead of inventing a plausible-sounding one.

How it actually works

Grounding connects generation to external evidence — retrieved documents, tool outputs, databases — so claims can be traced and verified. It’s the main practical defense against hallucination: not because the model gets smarter, but because it’s answering from provided facts rather than parametric memory alone.

For product teams

The difference between a demo and something you’d let talk to customers.

For engineers

Conditioning outputs on retrieved/verified external context and enabling attribution; the core mitigation for hallucination.

Related

Read anything AI without the jargon

Look up any term in plain English, or save terms as you read with the free Chrome extension.

Open DecoderAdd to Chrome