Glossary
AI hallucination (in analytics)
Definition
In analytics, an AI hallucination is a confident, plausible answer that is wrong, such as a metric calculated the wrong way or a number the data does not support.
What it means in practice
Analytics hallucinations are rarely invented facts. More often the AI guesses at a definition, joins the wrong tables or ignores a business rule, and presents the result with full confidence.
Because the answers look right, they are dangerous in decision-making. Grounding AI in certified business definitions and showing how each answer was calculated are the most effective defenses.
Example
An assistant reports 18% revenue growth by counting renewals as new bookings, a rule finance would never apply.