Hallucination
In one sentenceWhen AI confidently states something false: a fake statistic, a made-up quote, a person who does not exist.
What it means
LLMs generate the most plausible-sounding next words. Usually plausible and true line up. Sometimes they do not, and the model produces a fluent, confident answer that is simply wrong.
Hallucinations are most common with specific facts: numbers, dates, citations, quotes, URLs and details about less-famous people.
How to use it
- Never publish a statistic, quote or claim about a real person without checking the source.
- Ask the AI to cite sources, then click them. Tools with Grounding and Citations hallucinate far less.
- Add to your prompt: "If you are not sure, say so instead of guessing."
- Give the AI the facts yourself and ask it to write only from what you provided.
Prompt to try
Using ONLY the information below, write a short bio. If something is missing, write [CHECK] instead of inventing it. [paste notes]
Fill in the [brackets], then paste it into evyAI, ChatGPT, Claude or Gemini.
Related terms
Grounding Citations RAG Guardrails
Still fuzzy? Ask the evyAI agent to explain Hallucination with examples for your business.
Ask evyAI