Grounding
Tying AI outputs to verifiable sources and real evidence rather than letting the system make up information.
In Plain English
Grounding keeps AI systems honest by anchoring their answers in checkable facts. Without grounding, an AI might generate plausible-sounding answers that are completely false—a problem called 'hallucination' in the AI world. Grounded systems instead cite their sources: they retrieve information from trusted databases, web pages, or documents, then explain what they found and where. This matters enormously in high-stakes situations like medical advice, legal research, or financial decisions, where accuracy and accountability are essential. Grounding is like the difference between a news article with named sources versus gossip that sounds convincing but has no basis.
💡Real-World Example
Instead of a museum chatbot saying 'That painting was created in 1885,' a grounded system would say 'According to the museum's official database, that painting was created in 1885, by [artist name]'—and you could click through to verify the source yourself.
Related Terms
What did you think of our explanation?